EP3907997A1 - Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung - Google Patents

Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung Download PDF

Info

Publication number
EP3907997A1
EP3907997A1 EP21181798.6A EP21181798A EP3907997A1 EP 3907997 A1 EP3907997 A1 EP 3907997A1 EP 21181798 A EP21181798 A EP 21181798A EP 3907997 A1 EP3907997 A1 EP 3907997A1
Authority
EP
European Patent Office
Prior art keywords
block
prediction
prediction information
inter prediction
subject
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
EP21181798.6A
Other languages
English (en)
French (fr)
Inventor
Hiroya Nakamura
Shigeru Fukushima
Hideki Takehara
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
JVCKenwood Corp
Original Assignee
JVCKenwood Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by JVCKenwood Corp filed Critical JVCKenwood Corp
Publication of EP3907997A1 publication Critical patent/EP3907997A1/de
Pending legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • H04N19/513Processing of motion vectors
    • H04N19/517Processing of motion vectors by encoding
    • H04N19/52Processing of motion vectors by encoding by predictive encoding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103Selection of coding mode or of prediction mode
    • H04N19/105Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/119Adaptive subdivision aspects, e.g. subdivision of a picture into rectangular or non-rectangular coding blocks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/157Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
    • H04N19/159Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/46Embedding additional information in the video signal during the compression process
    • H04N19/463Embedding additional information in the video signal during the compression process by compressing encoding parameters before transmission
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • H04N19/583Motion compensation with overlapping blocks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/593Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques

Definitions

  • the present invention generally relates to a moving picture encoding and decoding technology, and more particularly, to a moving picture encoding and decoding technology utilizing motion compensation prediction.
  • MPEG-4 AVC/H.264 utilizes motion compensation, where a picture is partitioned into a plurality of rectangular blocks, a picture that has been already encoded or decoded is used as a reference picture, and a movement from the reference picture is predicted.
  • a method that predicts movement by this motion compensation is referred to as inter prediction or motion compensation prediction.
  • inter prediction of MPEG-4AVC/H.264 a plurality of pictures can be used as reference pictures, and a most appropriate reference picture is chosen from the plurality of reference pictures for each block so as to perform motion compensation prediction. Therefore, a reference index is assigned to each reference picture, and a reference picture is specified by the reference index.
  • L0 prediction L0 prediction
  • L1 prediction list 0 prediction
  • bi-prediction which uses two types of inter prediction (i.e., L0 prediction and L1 prediction).
  • L0 prediction and L1 prediction are also defined.
  • prediction for both directions is performed, inter-predicted signals of L0 prediction and L1 prediction respectively are multiplied by a weighting coefficient, an offset value is added and convolution is performed so as to generate an ultimate inter prediction picture signal.
  • the weighting coefficient used for weighted prediction and the offset value are defined by a typical value for each reference picture of each list on a picture by picture basis, and encoded.
  • Coding information relating to inter prediction includes: a prediction mode that differentiate among L0 prediction, L1 prediction, and bi-prediction for each block; a reference index that specifies a reference picture for each reference list of each block; and a motion vector that represents a direction of movement and an amount of movement of a block. This coding information is encoded and/or decoded.
  • a direct mode where inter prediction information of a block subject to encoding or decoding is generated from inter prediction information of a decoded block, is defined. Since encoding of inter prediction information is not required, encoding efficiency is improved with the direct mode.
  • a picture, of which the reference index of L1 is registered to 0, is used as a base picture colPic.
  • a block at a position identical to a block subject to encoding or decoding in the base picture colPic is used as a base block.
  • the motion vector of the base block L0 is used as a base motion vector mvCol. If the base block is not encoded by using L0 prediction but encoded by using L1 prediction, the motion vector of L1 of the base block is used as a base motion vector mvCol.
  • a picture referred to by the base motion vector mvCol is used as a reference picture of L0 in a time direct mode, and the base picture colPic is used as a reference picture of L1 in a time direct mode.
  • a motion vector mvLO of L0 and a motion vector mvL1 of L1 in a time direct mode are derived from a base motion vector mvCol by a scaling computation process.
  • POC POC of the reference picture of L0 in a time direct mode
  • tb POC of a picture subject to encoding or decoding ⁇ POC of a reference picture of L0 in a time direct mode
  • a motion vector mvLO of L0 in a time direct mode is derived from a base motion vector mvCol by a scaling computation process.
  • mvL 0 tb / td * mvCol
  • a motion vector mvL1 of L1 is derived by subtracting the base motion vector mvCol from the motion vector mvLO of L0 in a time direct mode.
  • mvL 1 mvL 0 ⁇ mvCol
  • Patent Document1 Japanese Patent Application Publication No. 2004-129191 . [SUMMARY OF THE INVENTION]
  • the degree of prediction accuracy is decreased and the encoding efficiency is not improved with a time direct mode.
  • the present inventors have recognized the necessity of further compression of coding information and reduction of the total amount of code in a moving picture encoding method that uses motion compensation prediction.
  • the present invention addresses the aforementioned issue, and a purpose thereof is to provide a moving picture encoding and decoding technology for reducing an amount of code of coding information by calculating a candidate of coding information, so as to improve the encoding efficiency.
  • a moving picture encoding device partitions a first block, which is obtained by partitioning each picture, into one or more second blocks and encodes a moving picture using inter prediction.
  • the device includes: a first prediction information derivation unit (130) configured to derive a first inter prediction information candidate from inter prediction information of a third block in a proximity of the second block subject to encoding in a picture subject to encoding; a second prediction information derivation unit (132) configured to derive a second inter prediction information candidate from inter prediction information of a fourth block at a position identical to or in a proximity of the second block subject to encoding in a picture different from the picture subject to encoding; a candidate list construction unit (133) configured to construct a prediction information candidate list including a prediction information candidate to which, in case that the first inter prediction information candidate is derived, the derived first inter prediction information candidate is added, and in case that the second inter prediction information candidate is derived, the derived second inter prediction information candidate is added, respectively; a selection unit (136) configured to select a candidate of inter prediction information to be used for the inter prediction of the second block subject to encoding from the one or more prediction information candidates in the prediction information candidate list, and configured to
  • a moving picture encoding device partitions a coding block into one or more prediction blocks in accordance with a partition mode so as to perform motion compensation prediction.
  • the device includes: a reference index derivation unit (131) configure to derive, in a merge mode where inter prediction information defined as a time merge candidate is derived from inter prediction information of a prediction block at a position identical to or in a proximity of a prediction block subject to encoding in a encoded picture different in time from the prediction block subject to encoding and inter prediction of the prediction block subject to encoding is executed on the basis of the derived inter prediction information, reference index information of a time merge candidate as reference index information that specifies a reference picture used in the prediction block subject to encoding; and a time merge candidate generating unit (132) configured to derive inter prediction information of time merge candidate based on the derived reference index information.
  • the reference index derivation unit (131) sets reference index information of a time merge candidate to a value of reference index information of an encoded prediction block adjacent to a left edge of the prediction block subject to encoding and outputs the reference index information
  • the partition mode is a mode where the coding block is divided by a vertical border into prediction blocks arranged horizontally
  • the reference index derivation unit (131) sets reference index information of a time merge candidate to a value of reference index information of an encoded prediction block adjacent to an upper edge of the prediction block subject to encoding and outputs the reference index information.
  • a moving picture encoding method partitions a first block, which is obtained by partitioning each picture, into one or more second blocks and encodes a moving picture using inter prediction.
  • the method includes: deriving a first inter prediction information candidate from inter prediction information of a third block in a proximity of the second block subject to encoding in a picture subject to encoding; deriving a second inter prediction information candidate from inter prediction information of a fourth block at a position identical to or in a proximity of the second block subject to encoding in a picture different from the picture subject to encoding; constructing a prediction information candidate list including a prediction information candidate to which, in case that the first inter prediction information candidate is derived, the derived first inter prediction information candidate is added, and in case that the second inter prediction information candidate is derived, the derived second inter prediction information candidate is added, respectively; selecting a candidate of inter prediction information to be used for the inter prediction of the second block subject to encoding from the one or more prediction information candidates
  • a moving picture decoding device decodes a bitstream where first block, which is obtained by partitioning each picture, is partitioned into one or more second blocks and a moving picture is encoded by inter prediction.
  • the device includes: a first prediction information derivation unit (230) configured to derive a first inter prediction information candidate from inter prediction information of a third block in a proximity of the second block subject to decoding in a picture subject to decoding; a second prediction information derivation unit (232) configured to derive a second inter prediction information candidate from inter prediction information of a fourth block at a position identical to or in a proximity of the second block subject to decoding in a picture different from the picture subject to decoding; a decoding unit (202) configured to decode an index indicating inter prediction information to be used for the inter prediction of the second block subject to decoding; a candidate list construction unit (233) configured to construct a prediction information candidate list including a prediction information candidate to which, in case that the first inter prediction information candidate is
  • a moving picture decoding device partitions a decoding block into one or more prediction blocks in accordance with a partition mode so as to perform motion compensation prediction.
  • the device includes: a reference index derivation unit (231) configure to derive, in a merge mode where inter prediction information defined as a time merge candidate is derived from inter prediction information of a prediction block at a position identical to or in a proximity of a prediction block subject to decoding in a decoded picture different in time from the prediction block subject to decoding and inter prediction of the prediction block subject to decoding is executed from the derived inter prediction information, reference index information of a time merge candidate as reference index information that specifies a reference picture used in the prediction block subject to decoding; and a time merge candidate generating unit configured to derive inter prediction information of time merge candidate based on the derived reference index information,.
  • the reference index derivation unit (231) sets reference index information of a time merge candidate to a value of reference index information of an decoded prediction block adjacent to a left edge of the prediction block subject to decoding and outputs the reference index information
  • the partition mode is a mode where the decoding block is divided by a vertical border into prediction blocks arranged horizontally
  • the reference index derivation unit (231) sets reference index information of a time merge candidate to a value of reference index information of an decoded prediction block adjacent to an upper edge of the prediction block subject to decoding and outputs the reference index information.
  • a moving picture decoding method decodes a bitstream where first block, which is obtained by partitioning each picture, is partitioned into one or more second blocks and a moving picture is encoded by inter prediction.
  • the method includes: deriving a first inter prediction information candidate from inter prediction information of a third block in a proximity of the second block subject to decoding in a picture subject to decoding; deriving a second inter prediction information candidate from inter prediction information of a fourth block at a position identical to or in a proximity of the second block subject to decoding in a picture different from the picture subject to decoding; decoding an index indicating inter prediction information to be used for the inter prediction of the second block subject to decoding; constructing a prediction information candidate list including a prediction information candidate to which, in case that the first inter prediction information candidate is derived, the derived first inter prediction information candidate is added, and in case that the second inter prediction information candidate is derived, the derived second inter prediction information candidate is added, respectively; and selecting a
  • an amount of generated code of coding information to be transmitted can be reduced and an encoding efficiency can be improved.
  • a plurality of motion vector predictors are derived from a motion vector of a block in the proximity of a block subject to encoding or from a motion vector of a block of an encoded picture, a difference vector between a motion vector of the block subject to encoding and a selected motion vector predictor is calculated, and the difference vector is encoded so as to reduce the amount of code.
  • coding information of a block in the proximity of a block subject to encoding or a block of an encoded picture is derived so as to reduce the amount of code.
  • a plurality of motion vector predictors are calculated from a motion vector of a block in the proximity of a block subject to decoding or from a motion vector of a block of a decoded picture
  • a motion vector of the block subject to decoding is calculated from a difference vector decoded from a bitstream and from a selected motion vector predictor so as to decode the block subject to decoding.
  • coding information of a block subject to decoding is derived.
  • a picture is partitioned in units of squares having a same arbitrary size as shown in Fig. 3 .
  • This unit is defined as a tree block and used as a basic unit of managing an address for specifying a block subject to encoding or decoding (Hereinafter, "a block subject to encoding or decoding" refers to a block subject to encoding in case of encoding process, and refers to a block subject to decoding in case of decoding process, unless otherwise noted.) in a picture.
  • a tree block comprises one brightness signal and two color-difference signals, except in case of a monochrome.
  • the size of a tree block can be determined freely in units of powers of 2 in accordance with the size of a picture and/or a texture in the picture.
  • a tree block can be hierarchically partitioned into four blocks (divided by two both vertically and horizontally) having a smaller block size, as necessary, by dividing a brightness signal and color-difference signals in the tree block. These blocks are defined as coding blocks respectively, and used as basic units in encoding and decoding.
  • a coding block also comprises one brightness signal and two color-difference signals, except in case of a monochrome.
  • the maximum size of a coding block is as same as the size of a tree block.
  • a coding block of minimum size is referred to as a minimum coding block, and the size thereof can be set freely in units of powers of 2.
  • a coding block A is defined by using a tree block as one coding block without partitioning the tree block.
  • a coding block B is a coding block made by partitioning a tree block into four blocks.
  • a coding block C is a coding block made by further partitioning a block, which is obtained by partitioning a tree block into four, into four blocks.
  • a coding block D is a coding block made by further partitioning a block, which is obtained by partitioning a tree block into four, into four blocks twice hierarchically.
  • the coding block D is a coding block of the minimum size.
  • a mode is switched between intra prediction (MODE_INTRA) and inter prediction (MODE_INTER) in units of coding blocks.
  • intra prediction prediction is performed from a surrounding encoded or decoded picture signals in a picture subject to encoding or decoding
  • decoded is used for a decoded picture that has been encoded in a encoding process, a decode prediction block that has been encoded in a encoding process, a decoded picture signal that has been encoded in a encoding process, or the like
  • decoded is used for a decoded picture, a decoded prediction block, a decoded picture signal, or the like.
  • This mode that distinguishes the intra prediction (MODE_INTRA) and the inter prediction (MODE_INTER) is defined as a prediction mode (PredMode).
  • the prediction mode (PredMode) has, as a value, the intra prediction (MODE_INTRA) or the inter prediction (MODE_INTER), and can select a type of prediction so as to perform encoding.
  • PartMode partition mode
  • PartMode partition mode
  • PartMode A partition mode (PartMode) where a brightness signal of a coding block is not partitioned and regarded as one prediction block as shown in Fig. 4A is defined as 2N by 2N partition (PART_2Nx2N).
  • Partition modes (PartMode) where a brightness signal of a coding block is partitioned into two prediction blocks arranged vertically as shown in Fig. 4B, 4C, and 4D are defined as 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), and 2N by nD partition (PART_2NxnD), respectively.
  • the 2N by N partition (PART_2NxN) is a partition mode where a block is partitioned with a ratio of 1:1, vertically
  • the 2N by nU partition (PART_2NxnU) is a partition mode where a block is partitioned with a ratio of 1:3, vertically
  • 2N by nD partition (PART_2NxnD) is a partition mode where a block is partitioned with a ratio of 3:1, vertically.
  • Partition modes (PartMode) where a brightness signal of a coding block is partitioned into two prediction blocks arranged horizontally as shown in Fig.
  • 4E, 4F, and 4G are defined as N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), and nR by 2N partition (PART_nRx2N), respectively.
  • the N by 2N partition (PART_Nx2N) is a partition mode where a block is partitioned with a ratio of 1:1, horizontally
  • the nL by 2N partition (PART_nLx2N) is a partition mode where a block is partitioned with a ratio of 1:3, horizontally
  • nR by 2N partition PART_nRx2N
  • a partition mode (PartMode) where a brightness signal of a coding block is partitioned vertically and horizontally into four blocks as shown in Fig. 4H is defined as N by N partition (PART_NxN).
  • PartMode For each partition mode (PartMode), a color-difference signal is also partitioned in a similar manner to that of horizontal and vertical partition ratio of a brightness signal.
  • a number starting from 0 is assigned to a prediction block in the coding block in the order of encoding. This number is defined as a partition index PartIdx.
  • a number written in each prediction block in the coding blocks shown in Fig. 4 indicates the partition index PartIdx of the prediction block.
  • the partition index PartIdx of an upper prediction block is set to 0 and the partition index PartIdx of a lower prediction block is set to 1.
  • the partition index PartIdx of a left prediction block is set to 0 and the partition index PartIdx of a right prediction block is set to 1.
  • the partition index PartIdx of a right prediction block is set to 1.
  • the partition index PartIdx of an upper left prediction block is set to 0, the partition index PartIdx of an upper right prediction block is set to 1, the partition index PartIdx of a lower left prediction block is set to 2, and the partition index PartIdx of a lower right prediction block is set to 3.
  • PredMode is inter prediction (MODE_INTER)
  • the 2N by 2N partition PART_2Nx2N
  • the 2N by N partition PART_2NxN
  • the 2N by nU partition PART_2NxnU
  • the 2N by nD partition PART_2NxnD
  • the N by 2N partition PART_Nx2N
  • the nL by 2N partition PART_nLx2N
  • nRx2N nR by 2N partition
  • the N by N partition can be defined as a partition mode (PartMode) in addition to the 2N by 2N partition (PART_2Nx2N), the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N by 2N partition (PART_Nx2N), the nL by 2N partition (PART_nLx2N), and the nR by 2N partition (PART_nRx2N).
  • PartMode partition mode in addition to the 2N by 2N partition (PART_2Nx2N), the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N by 2N partition (PART_Nx2N), the nL by 2N partition (PART_nLx2N), and the nR by 2N partition (PART_nRx
  • the prediction mode (PredMode) is intra prediction (MODE_INTRA)
  • PART_2Nx2N only the 2N by 2N partition (PART_2Nx2N) is defined as a partition mode (PartMode) for blocks except for the coding block D, which is the minimum coding block.
  • the N by N partition (PART_NxN) is defined as a partition mode (PartMode) in addition to the 2N by 2N partition (PART_2Nx2N).
  • the reason why the N by N partition (PART_NxN) is not defined except for the minimum coding block, is that coding block other than the minimum coding block can be partitioned into four blocks so as to represent smaller coding blocks.
  • each block such as a tree block, a coding block, a prediction block, a transformation block, or the like according to the present embodiment is represented by two dimensional coordinates where the position of the top left pixel of brightness signal in a screen of brightness signals is defined as the original point, and the position of the top left pixel of brightness signal included in the area of each block is represented by two dimensional coordinates (X, Y).
  • the positive directions are the horizontally-right direction and the vertically-down direction, respectively.
  • the unit of scale is pixel of brightness signal.
  • each block of color-difference signals is represented by coordinates of a pixel of a brightness signal included in an area of the block and the unit of scale is pixel, not only in case that a color-difference format is 4:4:4 where the picture sizes (the numbers of pixels) for brightness signals and color-difference signals are identical to each other, but also in case that the color-difference format is 4:2:0 or 4:2:2 where the picture sizes (the numbers of pixels) of brightness signals and color-difference signals are different.
  • a color-difference format is 4:4:4 where the picture sizes (the numbers of pixels) for brightness signals and color-difference signals are identical to each other, but also in case that the color-difference format is 4:2:0 or 4:2:2 where the picture sizes (the numbers of pixels) of brightness signals and color-difference signals are different.
  • a plurality of decoded pictures can be used as a reference picture in inter prediction where a prediction is made on the basis of picture signals of a decoded picture.
  • a reference index is assigned for each prediction block.
  • the inter prediction modes includes L0 prediction (Pred_L0), L1 prediction (Pred_L1), and bi-prediction (Pred_BI).
  • the reference picture is managed by L0 (reference list 0) and L1 (reference list 1), which are list structures, and a reference picture can be identified by specifying a reference index of L0 or L1.
  • the L0 prediction (Pred_L0) is inter prediction that refers to a reference picture managed by L0
  • the L1 prediction (Pred_L1) is inter prediction that refers to a reference picture managed by L1
  • the bi-prediction (Pred_BI) is prediction where both L0 prediction and L1 prediction are performed, and one reference picture managed by L0 and one reference picture managed by L1 are referred to.
  • inter prediction of a P slice only L0 prediction can be used.
  • L0 prediction, L1 prediction, and bi-prediction where L0 prediction and L1 prediction are averaged or added with weights, can be used.
  • a merge mode is not a mode that encode or decode inter prediction information (e.g., a prediction mode, a reference index, a motion vector, or the like) of a prediction block subject to encoding or decoding, but is a mode that performs inter prediction by deriving inter prediction information of a prediction block subject to encoding or decoding from inter prediction information of: a prediction block that is in a same picture with the prediction block subject to encoding or decoding and that is in the proximity of the prediction block subject to encoding or decoding; or a prediction block that is in a picture different in time from a picture of the prediction block subject to encoding or decoding, and that is at a position identical to or in the proximity (at a position in the proximity of) of the prediction block subject to encoding or decoding.
  • inter prediction information e.g., a prediction mode, a reference index, a motion vector, or the like
  • Each merge candidate is registered in a merge candidate list, and a merge candidate to be used for inter prediction is specified by a merge index.
  • Figs. 5 , 6 , 7 , and 8 illustrate: a deprivation of a spatial merge candidate; and a prediction block to be referred to when deriving a spatial merge candidate and when deriving the reference index of a time merge candidate, the prediction block being in a same picture with the prediction block subject to encoding or decoding and being in the proximity of the prediction block subject to encoding or decoding.
  • Fig. 9 illustrates a prediction block to be referred to when deriving the reference index of a time merge candidate, the prediction block being in an encoded or decoded picture different in time from the prediction block subject to encoding or decoding and being at a position identical to or in the proximity of the prediction block subject to encoding or decoding.
  • a prediction block A adjacent to the left edge of the prediction block subject to encoding or decoding, a prediction block B adjacent to the upper edge thereof, a prediction block C adjacent to the upper right vertex thereof, a prediction block D adjacent to the lower left vertex thereof, and a prediction block E adjacent to the upper left vertex thereof are defined as a prediction block in the spatial proximity.
  • a prediction block A10 at the lowest position among the plurality of prediction blocks adjacent to the left edge is defined as a prediction block A adjacent to the left edge according to the present embodiment.
  • a prediction block B10 at the right most position among the plurality of prediction blocks adjacent to the upper edge is defined as a prediction block B1 adjacent to the upper edge according to the present embodiment.
  • the prediction block A is defined as a prediction block A if the prediction block A is adjacent to the left edge of the prediction block subject to encoding or decoding, is defined as a prediction block D in case of being adjacent to the lower left vertex of the prediction block subject to encoding or decoding, and is defined as a prediction block E in case of being adjacent to the upper left vertex of the prediction block subject to encoding or decoding.
  • a prediction block A, a prediction block E, and a prediction block E are an identical prediction block.
  • the prediction block G is defined as a prediction block B if the prediction block G is adjacent to the upper edge of the prediction block subject to encoding or decoding, is defined as a prediction block C if the prediction block G is adjacent to the upper right vertex of the prediction block subject to encoding or decoding, and is defined as a prediction block E if the prediction block G is adjacent to the upper left vertex of the prediction block subject to encoding or decoding.
  • a prediction block B, a prediction block C, and a prediction block E are an identical prediction block.
  • an encoded or decoded prediction blocks T0 and T1 in a picture different in time from a picture of the prediction block subject to encoding or decoding and at a position identical to or in the proximity of the prediction block subject to encoding or decoding are defined as a prediction block at a same position different in time.
  • POC is defined as a parameter that is associated with a picture subject to encoding, and a value that increases one at a time in the output order of pictures is defined as PCD.
  • PCD a value that increases one at a time in the output order of pictures.
  • Fig. 1 is a block diagram showing a structure of a moving picture encoding device according to an embodiment of the invention.
  • the moving picture encoding device comprises a picture memory 101, a motion vector detection unit 102, a motion vector difference calculation unit 103, an inter prediction information derivation unit 104, a motion compensation prediction unit 105, an intra prediction unit 106, a prediction method determining unit 107, a residual signal generating unit 108, an orthogonal transformation/quantization unit 109, a first bitstream generating unit 110, a second bitstream generating unit 111, a multiplexing unit 112, an inverse quantization/inverse orthogonal transformation unit 113, a decoded picture signal convolution unit 114, a coding information storing memory 115, and a decoded picture memory 116.
  • the picture memory 101 temporarily stores a picture signal of a picture subject to encoding, the signal having been provided in the order of capturing and/or in the order of display time.
  • the picture memory 101 provides the stored picture signal of a picture subject to encoding, in units of predetermined pixel blocks, to the motion vector detection unit 102, the prediction method determining unit 107, and the residual signal generating unit 108.
  • the picture signals of pictures stored in the order of capturing and/or in the order of display time are rearranged in the order of encoding, and outputted from the picture memory 101 in units of pixel blocks.
  • the motion vector detection unit 102 detects motion vectors for respective prediction block size and respective prediction modes for each prediction block by block matching or the like between a picture signal provided from the picture memory 101 and a reference picture provided from the decoded picture memory 116.
  • the motion vector detection unit 102 provides the detected motion vector to the motion compensation prediction unit 105, the motion vector difference calculation unit 103, and the prediction method determining unit 107.
  • the motion vector difference calculation unit 103 calculates a plurality of candidates of a motion vector predictor and stores the candidates in a motion vector predictor list, selects an optimal motion vector predictor from the plurality of candidates of a motion vector predictor registered in the motion vector predictor list, calculates a motion vector difference from the motion vector detected by the motion vector detection unit 102 and from the motion vector predictor, and provides the calculated motion vector difference to the prediction method determining unit 107. Further, the motion vector difference calculation unit 103 provides the prediction method determining unit 107 with a motion vector predictor index that specifies a motion vector predictor selected from the candidates of a motion vector predictor registered in the motion vector predictor list.
  • the inter prediction information derivation unit 104 derives a merge candidate in a merge mode. By using coding information of a prediction block having already been encoded and stored in the coding information storing memory 115, the inter prediction information derivation unit 104 derives a plurality of merge candidates and registers the candidates in a merge candidate list, which will be described below, selects an appropriate merge candidate from the plurality of merge candidates registered in the merge candidate list, provides the motion compensation prediction unit 105 with inter prediction information (e.g., flags predFlagL0[xP][yP] and predFlagL1[xP][yP] indicating whether or not to use L0 prediction and/or L1 prediction, reference indexes refIdxL0[xP][yP] and refIdxL1[xP][yP], motion vectors mvL0[xP][yP] and mvL1[xP][yP], or the like) of each prediction block of the selected merge candidate, and provides the prediction
  • the motion compensation prediction unit 105 generates from a reference picture a prediction picture signal by inter prediction (motion compensation prediction) by using a motion vector detected by the motion vector detection unit 102 and the inter prediction information derivation unit 104, and provides the prediction picture signal to the prediction method determining unit 107.
  • inter prediction motion compensation prediction
  • L0 prediction and in L1 prediction one-directional prediction is performed.
  • Pred_BI bi-prediction
  • prediction for two directions are performed, respective inter-predicted signals for L0 prediction and L1 prediction are multiplied by a weighting coefficient adaptively, an offset value is added, and convolution is performed so as to generate an ultimate prediction picture signal.
  • the intra prediction unit 106 performs intra prediction for each intra prediction mode.
  • the intra prediction unit 106 generates a prediction picture signal by intra prediction from a decoded picture signal stored in the decoded picture memory 211, selects an appropriate intra prediction mode from a plurality of intra prediction modes, and provides the prediction method determining unit 107 with a selected intra prediction mode and a prediction picture signal according to the selected intra prediction mode.
  • the prediction method determining unit 107 determines a prediction mode PredMode and a partition mode PartMode that determine inter prediction (PRED_INTER) or intra prediction (PRED_INTRA) in units of optimal coding blocks from a plurality of prediction methods, by evaluating an amount of code of coding information and a residual signal, an amount of distortion between a prediction picture signal and a picture signal or the like for respective prediction methods.
  • PRED_INTER inter prediction
  • the prediction method determining unit 107 determines whether or not a mode is a merge mode in units of prediction blocks.
  • the prediction method determining unit 107 determines a merge index, and in case that the mode is not a merge mode, determines an inter prediction mode, a motion vector predictor index, reference indexes of L0 and L1, a motion vector difference, or the like, and provides the first bitstream generating unit 110 with coding information according to the determination.
  • the prediction method determining unit 107 stores information indicating the determined prediction method, and coding information including a motion vector or the like according to the determined prediction method in the coding information storing memory 115.
  • the coding information to be stored includes: a prediction mode PredMode and a partition mode PartMode of respective coding blocks; flags predFlagL0[xP][yP] and predFlagL1[xP][yP] indicating whether or not to use L0 prediction and L1 prediction of each prediction block; reference indexes refIdxL0[xP][yP] and refIdxL1[xP][yP] of L0 and L1; motion vectors mvL0[xP][yP] and mvL1[xP][yP] of L0 and L1; or the like.
  • xP and yP are indexes indicating the position of an upper left pixel in a prediction block in a picture.
  • the prediction mode PredMode is intra prediction (MODE_INTRA)
  • the prediction mode PredMode is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xP][yP] indicating whether or not to use L0 prediction is 1
  • the flag predFlagL1[xP][yP] indicating whether or not to use L1 prediction is 0.
  • the inter prediction mode is L1 prediction (Pred_L1)
  • the flag predFlagL0[xP][yP] indicating whether or not to use L0 prediction is 0
  • the flag predFlagL1[xP] [yP] indicating whether or not to use L1 prediction is 1.
  • the prediction method determining unit 107 provides the residual signal generating unit 108 and the decoded picture signal convolution unit 114 with a prediction picture signal according to the determined prediction mode.
  • the residual signal generating unit 108 generates a residual signal by performing subtraction between a picture signal subject to encoding and a prediction picture signal, and provides the residual signal to the orthogonal transformation/quantization unit 109.
  • the orthogonal transformation/quantization unit 109 performs orthogonal transformation and quantization on the residual signal in accordance with a quantization parameter so as to generate orthogonally-transformed and quantized residual signal and provides the orthogonally-transformed and quantized residual signal to the second bitstream generating unit 111 and to the inverse quantization/inverse orthogonal transformation unit 113. Further, the orthogonal transformation/quantization unit 109 stores the quantization parameter in the coding information storing memory 115.
  • the first bitstream generating unit 110 encodes coding information according to a prediction method determined for each coding block and each prediction block by the prediction method determining unit 107, in addition to information for each sequence, each picture, each slice, and each coding block. More specifically, the first bitstream generating unit 110 encodes coding information, such as, a prediction mode PredMode for each coding block, a partition mode PartMode, a flag determining whether or not a mode is a merge mode in case of inter prediction (PRED_INTER), a merge index in case of merge mode, an inter prediction mode in case that the mode is not a merge mode, a motion vector predictor index, information regarding a motion vector difference, or the like, in accordance with a prescribed syntax rule, which will be described later, so as to generate a first bitstream and provides the first bitstream to the multiplexing unit 112.
  • a prediction mode PredMode for each coding block such as, a prediction mode PredMode for each coding block, a partition mode PartMode,
  • the second bitstream generating unit 111 entropy encodes the orthogonally-transformed and quantized residual signal in accordance with a prescribed syntax rule so as to generate a second bitstream and provides the second bitstream to the multiplexing unit 112.
  • the first bitstream and the second bitstream are multiplexed in accordance with a prescribed syntax rule so as to generate a bitstream.
  • the inverse quantization/inverse orthogonal transformation unit 113 inverse quantizes and inverse-orthogonally transforms the orthogonally-transformed and quantized residual signal provided from the orthogonal transformation/quantization unit 109 so as to calculate a residual signal, and provide the residual signal to the decoded picture signal convolution unit 114.
  • the decoded picture signal convolution unit 114 convolves the prediction picture signal according to the determination by the prediction method determining unit 107 and the residual signal that is inverse-quantized and inverse-orthogonally-transformed in the inverse quantization/inverse orthogonal transformation unit 113 so as to generate a decoded picture and stores the decoded picture in the decoded picture memory 116.
  • a filtering process that reduces distortion such as block distortion or the like caused by coding is performed on the decoded picture and the decoded picture is stored in the decoded picture memory 116, accordingly.
  • Fig. 1 is a block diagram showing a structure of a moving picture decoding device according to an embodiment of the invention, the moving picture decoding device corresponding to the moving picture encoding device of Fig. 1 .
  • the moving picture decoding device according to the embodiment comprises a separation unit 201, a first bitstream decoding unit 202, a second bitstream decoding unit 203, a motion vector calculation unit 204, an inter prediction information derivation unit 205, a motion compensation prediction unit 206, an intra prediction unit 207, an inverse quantization/inverse orthogonal transformation unit 208, a decoded picture signal convolution unit 209, a coding information storing memory 210, and a decoded picture memory 211.
  • Decoding process of the moving picture decoding device of Fig. 2 corresponds to the decoding process provided in the moving picture encoding device of Fig. 1 . Therefore, respective configurations of the motion compensation prediction unit 206, the inverse quantization/inverse orthogonal transformation unit 208, the decoded picture signal convolution unit 209, the coding information storing memory 210, and the decoded picture memory 211 of Fig. 2 have a function corresponding to respective configurations of the motion compensation prediction unit 105, the inverse quantization/inverse orthogonal transformation unit 113, the decoded picture signal convolution unit 114, the coding information storing memory 115, and the decoded picture memory 116 of the moving picture encoding device of Fig. 1 .
  • the bitstream provided to the separation unit 201 is separated in accordance with a prescribed syntax rule.
  • the separated first bitstream is provided to the first bitstream decoding unit 202, and the second bitstream is provided to the second bitstream decoding unit 203.
  • the first bitstream decoding unit 202 decodes the provided bitstream and obtains information for each sequence, each picture, each slice, and each coding block, and coding information for each prediction block. More specifically, the first bitstream decoding unit 202 decodes coding information regarding, for example, a prediction mode PredMode that determines whether prediction is inter prediction (PRED_INTER) or intra prediction(PRED_INTRA) for each coding block, a partition mode PartMode, a flag determines whether or not a mode is a merge mode in case of inter prediction (PRED_INTER), a merge index in case that the mode is a merge mode, an inter prediction mode in case that the mode is not a merge mode, a motion vector predictor index, a motion vector difference, or the like, in accordance with a prescribed syntax rule, which will be described later, and provides the coding information to the motion vector calculation unit 204, the inter prediction information derivation unit 205 or the intra prediction unit 207.
  • a prediction mode PredMode that determines
  • the second bitstream decoding unit 203 decodes the provided bitstream so as to calculate an orthogonally-transformed and quantized residual signal and provides the orthogonally-transformed and quantized residual signal to the inverse quantization/inverse orthogonal transformation unit 208.
  • the motion vector calculation unit 204 derives, by using coding information of a picture signal already decoded and stored in the coding information storing memory 210, a plurality of candidates of a motion vector predictor and registers the candidates in a motion vector predictor list, which will be described later.
  • the motion vector calculation unit 204 selects a motion vector predictor according to a motion vector predictor index, which is decoded in the first bitstream decoding unit 202 and provided, from the plurality of candidates of a motion vector predictor registered in the motion vector predictor list, calculates a motion vector from the difference vector decoded in the first bitstream decoding unit 202 and from the selected motion vector predictor.
  • the motion vector calculation unit 204 provides the motion compensation prediction unit 206 with the motion vector with other coding information, and stores the motion vector and other coding information in the coding information storing memory 210.
  • the coding information to be provided and stored in this process includes: a prediction mode PredMode; partition mode PartMode; flags predFlagL0[xP][yP] and predFlagL1[xP][yP] indicating whether or not to use L0 prediction and L1 prediction respectively; reference indexes refIdxL0[xP][yP] and refIdxL1[xP][yP] of L0 and L1; motion vectors mvL0[xP][yP] and mvL1[xP][yP] of L0 and L1; or the like.
  • xP and yP are indexes indicating the position of an upper left pixel in a prediction block in a picture.
  • the flag predFlagLO indicating whether or not to use L0 prediction is 1
  • the flag predFlagL1 indicating whether or not to use L1 prediction is 0.
  • the flag predFlagLO indicating whether or not to use L0 prediction is 0, and the flag predFlagL1 indicating whether or not to use L1 prediction is 1.
  • a flag predFlagLO indicating whether or not to use L0 prediction and a flag predFlagL1[xP][yP] indicating whether or not to use L1 prediction are both 1.
  • the inter prediction information derivation unit 205 derives a merge candidate.
  • the inter prediction information derivation unit 205 derives a plurality of merge candidates and registers the candidates in a merge candidate list, which will be described later.
  • the inter prediction information derivation unit 205 selects, from the plurality of merge candidates registered in the merge candidate list, a merge candidate corresponding to a merge index that is decoded in the first bitstream decoding unit 202 and provided.
  • the inter prediction information derivation unit 205 provides the motion compensation prediction unit 206 with inter prediction information of the selected merge candidate such as, flags predFlagL0[xP][yP] and predFlagL1[xP][yP] indicating whether or not to use L0 prediction and L1 prediction, reference indexes refIdxL0[xP][yP] and refIdxL1[xP][yP] of L0 and L1, motion vectors mvL0[xP][yP] and mvL1[xP][yP] of L0 and L1, or the like, and stores the inter prediction information in the coding information storing memory 210.
  • xP and yP are indexes indicating the position of an upper left pixel in a prediction block in a picture. The detail on the configuration and operation of the inter prediction information derivation unit 205 will be given later.
  • the motion compensation prediction unit 206 generates a prediction picture signal by inter prediction (motion compensation prediction) by using inter prediction information calculated in the motion vector calculation unit 204 or in the inter prediction information derivation unit 205 from a reference picture stored in the decoded picture memory 211, and provides the prediction picture signal to the decoded picture signal convolution unit 209.
  • inter prediction motion compensation prediction
  • two motion compensation prediction picture signals of L0 prediction and L1 prediction are multiplied by a weighting coefficient adaptively, and convolution is performed so as to generate an ultimate prediction picture signal.
  • the intra prediction unit 207 performs intra prediction when the prediction mode PredMode of a prediction block to be decoded is intra prediction (PRED_INTRA).
  • the coding information decoded in the first bitstream decoding unit includes an intra prediction mode. Based on the intra prediction mode, a prediction picture signal is generated by intra prediction from the decoded picture signal stored in the decoded picture memory 211, and provided to the decoded picture signal convolution unit 209. Both flags, predFlagL0[xP][yP] and predFlagL1[xP][yP] indicating whether or not to use L0 prediction and L1 prediction, are set to 0 and stored in the coding information storing memory 210.
  • xP and yP are indexes indicating the position of an upper left pixel in a prediction block in a picture.
  • the inverse quantization/inverse orthogonal transformation unit 208 inverse-orthogonally transforms and inverse quantizes the orthogonally-transformed and quantized residual signal, which is decoded in the first bitstream decoding unit 202, and obtains an inverse-orthogonally transformed and inverse quantized residual signal.
  • the decoded picture signal convolution unit 209 decodes a decoded picture signal by convolving the prediction picture signal that is inter predicted in the motion compensation prediction unit 206 or the prediction picture signal that is intra predicted in the intra prediction unit 207 and the residual signal that is inverse-orthogonally-transformed and inverse-quantized by the inverse quantization/inverse orthogonal transformation unit 208, and stores the decoded picture signal in the decoded picture memory 211.
  • a filtering process that reduces distortion such as block distortion or the like caused by coding is performed on the decoded picture and the decoded picture is stored in the decoded picture memory 211,depending on circumstances.
  • Fig. 10 shows a first syntax structure that is described in a slice header for each slice of a bitstream generated in accordance with the present embodiment.
  • the figure shows only syntax elements related to the present embodiment.
  • a slice type is a B slice
  • a flag collocated_from_l0_flag is established.
  • the flag collocated_from_l0_flag indicates which of the reference pictures a picture colPic of a different time uses, namely, a reference picture registered in a reference list of L0 or a reference picture registered in a reference list of L1 of a picture including a prediction block subject to processing.
  • the picture colPic of a different time is used when deriving a candidate of a motion vector predictor over time or when deriving a merge candidate.
  • the detail on the flag collocated_from_l0_flag will be described later.
  • the syntax elements described above may be established in a picture parameter set that describes a syntax element that is set for each picture.
  • Fig. 11 shows a syntax pattern described for each prediction block.
  • the value of the prediction mode PredMode of a prediction block is inter prediction (MODE_INTER)
  • merge_flag[x0][y0] that indicates whether or not a mode is a merge mode
  • x0 and y0 are indexes indicating the position of an upper left pixel in a prediction block in a picture of brightness signals
  • merge_flag[x0][y0] is a flag indicating whether or not the mode of a prediction block positioned at (x0, y0) in the picture is a merge mode or not.
  • merge_flag[x0][y0] 1, which indicates that the mode is a merge mode
  • a syntax element merge_idx[x0][y0] of an index of a merge list which is a list of merge candidates to be referred to, is established.
  • x0 and y0 are indexes indicating the position of an upper left pixel in a prediction block in a picture
  • merge_flag[x0][y0] is a merge index of a prediction block positioned at (x0, y0) in the picture.
  • Fig. 12 shows an example of an entropy code of a syntax element merge_idx[x0][y0] of a merge index.
  • the number of candidates of merge is set to 5.
  • the code of syntax element merge_idx[x0][y0] of a merge index is '0', '10', '110', '1110', and '1111', respectively.
  • merge_flag[x0][y0] indicates that the mode is not a merge mode
  • the slice type is a B slice
  • L0 prediction Pred_L0
  • L1 prediction Pred_L1
  • Pred_BI bi-prediction
  • syntax elements of reference index ref_idx_l0[x0][y0] and ref_idx_l1[x0][y0] for specifying a reference picture and syntax elements mvd_l0[x0][y0][j] and mvd_l1[x0][y0][j] of a motion vector difference, which is a difference between a motion vector of a prediction block derived by detection of a motion vector and a motion vector predictor, are established.
  • x0 and y0 are indexes indicating the position of an upper left pixel in a prediction block in a picture.
  • ref_idx_l0[x0][y0] and mvd_l0[x0][y0][j] are a reference index and a motion vector difference of L0 of a prediction block positioned at (x0, y0) in the picture, respectively.
  • ref_idx_l1[x0][y0],mvd_l1[x0][y0][j] are a reference index and a motion vector difference of L1 of a prediction block positioned at (x0, y0) in the picture, respectively.
  • j indicates components of motion vector difference.
  • a value 0 for j represents x component
  • a value 1 for j represents y component.
  • syntax elements mvp_idx_l0[x0][y0] and mvp_idx_l1[x0][y0] of index of a motion vector predictor list which is a list of candidates of motion vector predictor to be referred to, is established.
  • x0 and y0 are indexes indicating the position of an upper left pixel in a prediction block in a picture.
  • mvp_idx_l0[x0][y0] and mvp_idx_l1[x0][y0] are motion vector predictor index of L0 and L1 of a prediction block positioned at (x0, y0) in the picture.
  • the value of the number of candidates is set to 2.
  • a method for deriving inter prediction information according to the embodiment is performed in the inter prediction information derivation unit 104 of the moving picture encoding device of Fig. 1 and in the inter prediction information derivation unit 205 of the moving picture decoding device of Fig. 2 .
  • the method for deriving inter prediction information is executed for each prediction block that constitutes a coding block at any of the encoding processes or decoding processes. If a prediction mode PredMode of a prediction block is inter prediction (MODE_INTER) and a mode is a merge mode, and in case of encoding, the method is used when deriving a prediction mode, a reference index, and a motion vector of a prediction block subject to encoding by using a prediction mode, a reference index, and a motion vector of an encoded prediction block.
  • a prediction mode PredMode of a prediction block is inter prediction (MODE_INTER) and a mode is a merge mode
  • the method is used when deriving a prediction mode, a reference index, and a motion vector of a prediction block subject to decoding by using a prediction mode, a reference index, and a motion vector of a decoded prediction block.
  • a merge candidate is derived not only from the five prediction blocks (i.e., a prediction block A adjacent to the left, a prediction block B adjacent to the top, a prediction block C adjacent to the upper right, a prediction block D adjacent to the lower left, and a prediction block E adjacent to the upper left) explained while referring to Figs. 5 , 6 , 7 , and 8 , but also from a prediction block of a prediction block Col (T0 or T1) at an identical position or in a proximity and different in time, which has been explained while referring to Fig. 9 .
  • the inter prediction information derivation unit 104 of the moving picture encoding device and the inter prediction information derivation unit 205 of the moving picture decoding device register those merge candidates in merge candidate list in prescribed order that is common between encoding and decoding.
  • the inter prediction information derivation unit 104 of the moving picture encoding device determines a merge index that specifies an element of a candidate list and encodes via a first bitstream generating unit.
  • the inter prediction information derivation unit 205 of the moving picture decoding device selects a prediction block from the merge candidate list in accordance with a merge index decoded in the first bitstream decoding unit 202 and provided, and performs motion compensation prediction by using inter prediction information (e.g., a prediction mode, a reference index, a motion vector, or the like) of the selected merge candidate.
  • inter prediction information e.g., a prediction mode, a reference index, a motion vector, or the like
  • Fig. 13 shows a detailed configuration of the inter prediction information derivation unit 104 of the moving picture encoding device of Fig. 1 .
  • Fig. 14 shows a detailed configuration of the inter prediction information derivation unit 205 of the moving picture decoding device of Fig. 2 .
  • Portions surrounded by a thick line box in Figs. 13 and 14 indicate the inter prediction information derivation unit 104 of the moving picture encoding device and the inter prediction information derivation unit 205 of the moving picture decoding device, respectively.
  • portions surrounded by a thick dotted line in the boxes indicate a construction unit 120 of merge candidate list of the moving picture encoding device and a construction unit 220 of merge candidate list of the moving picture decoding device, respectively.
  • the construction unit of merge candidate list is provided also in the moving picture decoding device that corresponds to the moving picture encoding device according to the embodiment so that same derivation results that do not contradict between encoding and decoding can be obtained.
  • merge candidate derivation and merge candidate list construction process in merge candidate derivation and in merge candidate list construction process in the construction unit 120 of merge candidate list of the moving picture encoding device and the construction unit 220 of merge candidate list of the moving picture decoding device, it is assumed that merge candidate derivation and merge candidate list construction process of a prediction block subject to processing are performed without referring to a prediction block included in a coding block identical to a coding block including a prediction block subject to processing.
  • Fig. 15 illustrates a prediction block that is in the proximity of a prediction block subject to processing for each partition mode (PartMode) of the prediction block subject to processing.
  • A0, B0, C0, D0, and E0 indicates: a prediction block A adjacent to the left edge of the prediction block subject to processing of which the partition index PartIdx is 0; a prediction block B adjacent to the upper edge thereof; a prediction block C adjacent to the upper right vertex thereof; a prediction block D adjacent to the lower left vertex thereof; and a prediction block E adjacent to the upper left vertex thereof, respectively.
  • A1, B1, C1, D1, and E1 indicates: a prediction block A adjacent to the left edge of the prediction block subject to processing of which the partition index PartIdx is 1; a prediction block B adjacent to the upper edge thereof; a prediction block C adjacent to the upper right vertex thereof; a prediction block D adjacent to the lower left vertex thereof; and a prediction block E adjacent to the upper left vertex thereof, respectively.
  • A2, B2, C2, D2, and E2 indicates: a prediction block A adjacent to the left edge of the prediction block subject to processing of which the partition index PartIdx is 2; a prediction block B adjacent to the upper edge thereof; a prediction block C adjacent to the upper right vertex thereof; a prediction block D adjacent to the lower left vertex thereof; and a prediction block E adjacent to the upper left vertex thereof, respectively.
  • A3, B3, C3, D3, and E3 indicates: a prediction block A adjacent to the left edge of the prediction block subject to processing of which the partition index PartIdx is 3; a prediction block B adjacent to the upper edge thereof; a prediction block C adjacent to the upper right vertex thereof; a prediction block D adjacent to the lower left vertex thereof; and a prediction block E adjacent to the upper left vertex thereof, respectively.
  • Figs. 15B, 15C, and 15D show prediction blocks in the proximity in case that a partition mode (PartMode) is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), or 2N by nD partition (PART_2NxnD), where a coding block subject to processing is partitioned into two prediction blocks arranged vertically.
  • the prediction block B1 in the proximity of a prediction block subject to processing of which PartIdx is 1, is a prediction block of which PartIdx is 0.
  • merge candidate derivation and merge candidate list construction of a prediction block, of which PartIdx is 1 by referring to the prediction block B1 cannot be processed until merge candidate derivation and merge candidate list construction process of a prediction block, which is a prediction block B1 and belongs to a same coding block and of which PartIdx is 0, is completed and a merge candidate to be used is specified.
  • a partition mode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), or 2N by nD partition (PART_2NxnD)
  • merge candidate derivation and merge candidate list construction process of two prediction blocks in the coding block can be processed in parallel by referring to coding information of the prediction block A1, C1, D1, or E1, which are not included in a coding block identical to a coding block including the prediction block subject to processing, and by performing merge candidate derivation and merge candidate list construction process of a prediction block of which PartIdx is 1, without referring to coding information of the prediction block B1, which is a prediction block adjacent to the upper edge of the prediction block subject to processing, and is included in a coding block identical to a coding block including the prediction block subject to processing, and of which PartIdx is 0.
  • Figs. 15F, 15F, and 15G show prediction blocks in the proximity in case that a partition mode (PartMode) is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where a coding block subject to processing is partitioned into two prediction blocks arranged horizontally.
  • the prediction block A1 in the proximity of a prediction block subject to processing, of which PartIdx is 1, is a prediction block of which PartIdx is 0.
  • merge candidate derivation and merge candidate list construction of a prediction block, of which PartIdx is 1 by referring to the prediction block A1 cannot be processed until merge candidate derivation and merge candidate list construction process of a prediction block, which is a prediction block A1 and belongs to a same coding block and of which PartIdx is 0, is completed and a merge candidate to be used is specified.
  • a partition mode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), and nR by 2N partition (PART_nRx2N)
  • merge candidate derivation and merge candidate list construction process of respective prediction blocks in the coding block can be processed in parallel by referring to coding information of the prediction block B1, C1, D1, or E1, which are not included in a coding block identical to a coding block including the prediction block subject to processing, and by performing merge candidate derivation and merge candidate list construction process of a prediction block of which PartIdx is 1, without referring to coding information of the prediction block A1, which is a prediction block adjacent to the left edge of the prediction block subject to processing, and is included in a coding block identical to a coding block including the prediction block subject to processing, and of which PartIdx is 0.
  • FIG. 16 shows a flowchart illustrating merge candidate derivation and merge candidate list construction process by sequential processing
  • Fig. 17 shows a flowchart illustrating merge candidate derivation and merge candidate list construction process by parallel processing.
  • merge candidate is derived and merge candidate list is constructed for a prediction block, of which a partition index PartIdx is 0, of a coding block subject to processing (step S101).
  • partition mode PartMode
  • PART_2Nx2N 2N partition
  • partition mode (PartMode) is not 2N by 2N partition (PART_2Nx2N) (NO in step S102), that is, in case that the partition mode (PartMode) is the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N by 2N partition (PART_Nx2N), the nL by 2N partition (PART_nLx2N), or the nR by 2N partition (PART_nRx2N), merge candidate is derived and merge candidate list is constructed for a prediction block, of which a partition index PartIdx is 1, of a coding block subject to processing (step S103), and the merge candidate derivation and merge candidate list construction process is completed, accordingly.
  • the partition mode (PartMode) is the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition
  • merge candidate is derived and merge candidate list is constructed for a prediction block, of which a partition index PartIdx is 0, of a coding block subject to processing (step S101).
  • step S102 in case that the partition mode (PartMode) is not 2N by 2N partition (PART_2Nx2N) (NO in step S102), that is, in case that the partition mode (PartMode) is the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N by 2N partition (PART_Nx2N), the nL by 2N partition (PART_nLx2N), or the nR by 2N partition (PART_nRx2N), merge candidate is derived and merge candidate list is constructed for a prediction block, of which a partition index PartIdx is 1, of a coding block subject to processing (step S103).
  • PartMode is the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N by 2N partition (PART_Nx2N), the nL by 2N partition (PART_nLx2N), or the nR by 2N partition (PART_nRx2N),merge candidate derivation and merge candidate list construction process of two prediction blocks, of which a partition index PartIdx are 0 and 1 respectively, can be started simultaneously, by performing merge candidate derivation and merge candidate list construction for a prediction block, of which a partition index PartIdx is 1, without referring to a prediction block, of which a partition index PartIdx is 0, in the same coding block.
  • PartMode is the 2N by N partition (PART_2NxN), the 2N by nU partition (PART_2NxnU), the 2N by nD partition (PART_2NxnD), the N
  • merge candidate derivation and merge candidate list construction process of a prediction block, of which a partition index PartIdx is 1, can be performed without referring to a prediction block, of which a partition index PartIdx is 0.
  • N by N partition (PART_NxN) is not defined as a partition mode (PartMode) according to the present embodiment, it is also possible to define the N by N partition (PART_NxN).
  • Fig. 15H shows prediction blocks in the proximity in case that the partition mode (PartMode) is N by N partition (PART_NxN) in which a brightness signal of a coding block subject to processing is partitioned into four blocks vertically and horizontally.
  • the prediction block B1 in the proximity of a prediction block subject to processing, of which PartIdx is 1, is a prediction block of which PartIdx is 0.
  • merge candidate derivation and merge candidate list construction of a prediction block, of which PartIdx is 1 by referring to the prediction block A1 cannot be processed until merge candidate derivation and merge candidate list construction process of a prediction block, which is a prediction block A1 and belongs to a same coding block and of which PartIdx is 0, is completed and a merge candidate to be used is specified.
  • a partition mode is N by N partition (PART_NxN)
  • PartIdx of a prediction block subject to processing 1
  • merge candidate derivation and merge candidate list construction process of respective prediction blocks in the coding block can be processed in parallel by performing merge candidate derivation and merge candidate list construction process of a prediction block of which PartIdx is 1, without referring to coding information of the prediction block A1, of which PartIdx is 0.
  • the prediction block B2 in the proximity of a prediction block subject to processing of which PartIdx is 2 is a prediction block of which PartIdx is 0, and the prediction block C2 is a prediction block of which PartIdx is 1.
  • merge candidate derivation and merge candidate list construction of a prediction block, of which PartIdx is 2 by referring to the prediction blocks B2 and C2 cannot be processed until merge candidate derivation and merge candidate list construction process of prediction blocks, which are prediction blocks B2 and C2 and belong to the same coding block and of which PartIdx are 0 and 1 respectively, is completed and a merge candidate to be used is specified.
  • a partition mode is N by N partition (PART_NxN)
  • PartIdx of a prediction block subject to processing is 2
  • merge candidate derivation and merge candidate list construction process of respective prediction blocks in the coding block can be processed in parallel by performing merge candidate derivation and merge candidate list construction process of a prediction block of which PartIdx is 2, without referring to coding information of the prediction blocks B2 and C2, of which PartIdx are 0 and 1, respectively.
  • a partition mode is N by N partition (PART_NxN)
  • PartIdx of a prediction block subject to processing is 3
  • merge candidate derivation and merge candidate list construction process of respective prediction blocks in the coding block can be processed in parallel by performing merge candidate derivation and merge candidate list construction process of a prediction block of which PartIdx is 3, without referring to coding information of the prediction blocks E3, B3, and A3, of which PartIdx are 0 , 1, and 2, respectively.
  • the inter prediction information derivation unit 104 in Fig. 13 includes a spatial merge candidate generating unit 130, a reference index derivation unit 131, a time merge candidate generating unit 132, a merge candidate registering unit 133, a merge candidate identification determining unit 134, a merge candidate supplementing unit 135, and a coding information selection unit 136.
  • the inter prediction information derivation unit 205 in Fig. 14 includes a spatial merge candidate generating unit 230, a reference index derivation unit 231 of time merge candidate, a time merge candidate generating unit 232, a merge candidate registering unit 233, a merge candidate identification determining unit 234, a merge candidate supplementing unit 235, and a coding information selection unit 236.
  • Fig. 18 shows a flowchart illustrating a procedure of a merge candidate derivation and merge candidate list construction process, that is a function common to the inter prediction information derivation unit 104 of the moving picture encoding device and the inter prediction information derivation unit 205 of the moving picture decoding device according to an embodiment of the invention.
  • An explanation on the processes will be given below sequentially.
  • an explanation where a slice type slice_type is a B slice is given in the following explanation, the explanation can be also applied to the case of P slice, unless otherwise noted.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • spatial merge candidates A, B, C, D, and E from respective prediction blocks A, B, C, D, and E in the proximity of a block subject to encoding or decoding are derived and outputted.
  • N that indicates A, B, C, D, E, or a spatial merge candidate Col is defined.
  • a merge candidate is derived without referring to a prediction block included in a coding block identical to the coding block that includes a prediction block subject to processing
  • a spatial merge candidate included in the coding block identical to the coding block including a prediction block subject to processing is not derived.
  • a reference index of a time merge candidate is derived from a prediction block in the proximity of a block subject to encoding or decoding are derived and output (step S202).
  • a reference index of a time merge candidate is derived without referring to a prediction block included in a coding block identical to the coding block that includes a prediction block subject to processing.
  • a slice type slice_type is a P slice and inter prediction is performed by using inter prediction information of the time merge candidate
  • the reference index of L0 is derived in order to perform L0 prediction (Pred_L0).
  • respective reference indexes of L0 and L1 are derived in order to perform bi-prediction (Pred_BI).
  • a time merge candidate is derived from a picture of a different time and the time merge candidate is outputted.
  • a merge candidate list is generated, and spatial merge candidates A, B, C, D, and E, and a time merge candidate Col are added to the merge candidate list mergeCandList, by which, the merge candidate list mergeCandList is established, and the merge candidate list mergeCandList is outputted (step S204).
  • a merge candidate list is generated, in case that motion vectors of reference indexes of an identical merge candidate have a same value in the merge candidate list mergeCandList, the merge candidate is omitted except for a merge candidate that is earliest in the order, and the merge candidate list mergeCandList is outputted (step S205).
  • a merge candidate list is supplemented so that the number of merge candidates registered in the merge candidate list mergeCandList, and the merge candidate list mergeCandList is outputted (step S206).
  • the number of merge candidates is set to 5.
  • a merge candidate of which the prediction mode is bi-prediction (Pred_BI) and which is made by changing the combination of L0 prediction and L1 prediction between merge candidates that have been registered already; a merge candidate of which the prediction mode is bi-prediction (Pred_BI), which has a different reference index, and the value of the motion vector of which is (0, 0); or the like is added.
  • a method for deriving a merge candidate N from a prediction block N in the proximity of a prediction block subject to encoding or decoding which is a processing procedure of step S201 of Fig. 18 .
  • Fig. 19 shows a flowchart illustrating procedure of spatial merge candidate derivation process of step S201 in Fig. 18 .
  • N represents A (left), B (up), C (upper right), D(lower left), or E (upper left) indicating an are of a prediction block in the proximity.
  • up to four spatial merge candidates are derived from five prediction blocks in the proximity.
  • Fig. 18 by assigning A as a parameter N, coding information of a prediction block A adjacent to the left edge of a prediction block subject to encoding or decoding is checked and a merge candidate A is derived, by assigning B as a parameter N, coding information of a prediction block B adjacent to the upper edge is checked and a merge candidate B is derived, by assigning C as a parameter N, coding information of a prediction block C adjacent to the upper right is checked and a merge candidate C is derived, by assigning D as a parameter N, coding information of a prediction block D adjacent to the lower left is checked and a merge candidate D is derived, and by assigning E as a parameter N, coding information of a prediction block E adjacent to the upper left is checked and a merge candidate E is derived(step S1101 - step S1114).
  • a flag of a merge candidate E availableFlagE is set to 0 (step S1107), both values of motion vectors mvL0E and mvL1E of merge candidate E are set to (0,0) (step S1108), both values of flags predFlagL0E and predFlagLlE of merge candidate E are set to 0 (step S1109), and this spatial merge candidate derivation process is completed.
  • up to four spatial merge candidate are derived from prediction blocks in the proximity. Therefore, in case that four spatial merge candidates have been derived already, it is not required to perform the derivation process of spatial merge candidates any more.
  • step S1103 in case that the parameter N is not E, or in case that values of a flags availableFlagA, availableFlagB, availableFlagC, and availableFlagD are added and the total thereof is not 4 (NO in S1102), the method proceeds to step S1103.
  • the prediction block N in the proximity is included in a coding block identical to a coding block including a prediction block subject to derivation (YES in step S1103)
  • the value of a flag availableFlagN of a merge candidate N is set to 0 (step S1107)
  • both values of motion vectors mvLON and mvL1N of the merge candidate N are set to (0,0) (step S1108)
  • both values of flags predFlagLON and predFlagL1N of the merge candidate N are set to 0 (step S1109).
  • the value of a flag availableFlagN of merge candidate N is set to 0 (step S1107), by not referring to coding information of the prediction block N in the proximity and by setting the value of flag availableFlagN of the merge candidate N to 0, so as not to using as a spatial merge candidate, merge candidate derivation and merge candidate list construction process for respective prediction blocks in a same coding block can be processed in parallel.
  • the case of prediction block B adjacent to the upper edge of a prediction block subject to derivation where a partition mode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), or 2N by nD partition (PART_2NxnD) and where that PartIdx of a prediction block subject to processing is 1, corresponds to the case where the prediction block N in the proximity is included in a coding block identical to a coding block including a prediction block subject to derivation.
  • the prediction block B adjacent to the upper edge of a prediction block subject to derivation is a prediction block which is included in a coding block identical to a coding block including a prediction block subject to derivation and of which the PartIdx is 0. Therefore, by not referring to coding information of the prediction block B in the proximity and by setting the value of flag availableFlagB of the merge candidate B to 0 so as not to using as a spatial merge candidate, merge candidate derivation and merge candidate list construction process for respective prediction blocks in a same coding block can be processed in parallel.
  • a partition mode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and where PartIdx of a prediction block subject to processing is 1, also corresponds to the case where the prediction block N in the proximity is included in a coding block identical to a coding block including a prediction block subject to derivation.
  • the prediction block A adjacent to the left edge of a prediction block subject to derivation is a prediction block which is included in a coding block identical to a coding block including a prediction block subject to derivation and of which the PartIdx is 0. Therefore, by not referring to coding information of the prediction block A in the proximity and by setting the value of flag availableFlagA of the merge candidate A to 0 so as not to using as a spatial merge candidate, merge candidate derivation and merge candidate list construction process for respective prediction blocks in a same coding block can be processed in parallel.
  • PartMode is N by N partition (PART_NxN)
  • PartIdx of a prediction block subject to processing is 1, 2, or 3
  • the prediction block N in the proximity is sometimes included in a coding block identical to a coding block including a prediction block subject to derivation.
  • the prediction block N in the proximity is not included in a coding block identical to a coding block including a prediction block subject to derivation (YES in step S1103)
  • the prediction block N in the proximity of a prediction block subject to encoding or decoding is specified, and in case that respective prediction blocks N are available, coding information of a prediction block N is derived from the coding information storing memory 115 or from the coding information storing memory 210 (step S1104).
  • step S1106 If the prediction block N in the proximity cannot be used (NO in step S1105), or the prediction mode PredMode of the prediction block N is intra prediction (MODE_INTRA) (YES in step S1106), the value of a flag availableFlagN of a merge candidate N is set to 0 (step S1107), both values of motion vectors mvLON and mvL1N of the merge candidate N are set to (0,0) (step S1108), and both values of flags predFlagLON and predFlagL1N of the merge candidate N are set to 0 (step S1109).
  • the case where a prediction block N in the proximity cannot be used corresponds, for example, to a case where the prediction block N in the proximity is positioned outside of a slice subject to encoding or decoding, a case where encoding or decoding process has not been completed because the block N is late in the processing order of encoding or decoding process, etc.
  • the prediction block N in the proximity is outside of a coding block identical to a coding block including a prediction block subject to derivation (YES in step S1104), the prediction block N in the proximity is available (YES in step S1105), and in case that the prediction mode PredMode of the prediction block N is not an intra prediction (MODE_INTRA) (YES in step S1106), the inter prediction information of the prediction block N is used as the inter prediction information of the merge candidate N.
  • the prediction mode PredMode of the prediction block N is not an intra prediction (MODE_INTRA)
  • the value of a flag availableFlagN of the merge candidate N is set to 1 (step S1110), the motion vectors mvLON and mvL1N of the merge candidate N are set to a value identical to that of the motion vectors mvL0N[xN][yN] and mvL1N[xN][yN] of respective prediction block N (step S1111), the reference indexes refIdxL0N and refIdxL1N of the merge candidate N are set to a value identical to that of the reference indexes refIdxL0[xN][yN] and refIdxL1[xN][yN] of respective prediction block N (step S1112), and flags predFlagLON and predFlagL1N of a merge candidate N are set to flags predFlagL0[xN][yN] and predFlagL1[xN][yN] of respective prediction block N (step S1113).
  • reference indexes of a time merge candidate is derived by using a reference index of a spatial merge candidate, that is, a reference index used in a prediction block in the proximity of a prediction block subject to encoding or decoding. This is because, in case that a time merge candidate is selected, a reference index of a prediction block subject to encoding or decoding has a high correlation with the reference index of a prediction block in the proximity of a prediction block subject to encoding or decoding, which will be a spatial merge candidate.
  • a reference index of a prediction block A adjacent to the left edge of the prediction block subject to derivation, or a prediction block B adjacent to the upper edge thereof is used, except for a sixth exemplary embodiment and a seventh exemplary embodiment, which will be described later.
  • the prediction blocks A and B adjacent to the edge of the prediction block subject to encoding or decoding, among the prediction blocks A,B,C,D, and E in the proximity, which are also a spatial merge candidates have a higher correlation than the prediction blocks C,D, and E adjacent only to a vertex.
  • an improvement effect of encoding efficiency can be obtained by deriving a reference index of a time merge candidate, and the amount of calculation and a frequency of access to memory relating to derivation process of a reference index of a time merge candidate can be reduced.
  • Fig. 20 shows an adjacent block referred to in the derivation process of a reference index of a time merge candidate according to the first exemplary embodiment of the present embodiment.
  • PartMode partition mode
  • which of prediction blocks: a prediction block adjacent to the left edge of a prediction block subject to derivation, or a prediction block adjacent to the upper edge thereof is used is switched.
  • a prediction block adjacent to the edge outside of a coding block is referred to.
  • PartMode is 2N by 2N partition (PART_2Nx2N)
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition(PART_2NxnU), or 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, as shown in Fig. 20B, 20C, and 20D , prediction blocks adjacent to the left edge of respective prediction blocks subject to derivation are referred to, and respective reference indexes of LX of time merge candidates are set to values of reference indexes of LX of prediction blocks adjacent to the left edge of respective prediction blocks subject to derivation.
  • a prediction block A0 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • a prediction block A1 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block A1. Since both prediction blocks A0 and A1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where coding block subject to processing is partitioned into two prediction blocks arranged horizontally, as shown in Fig. 20E, 20F, and 20G , a prediction block adjacent to the upper edge of a prediction block subject to derivation is referred to, and respective reference indexes of LX of time merge candidates are set to values of reference indexes of LX of prediction blocks adjacent to the upper edge of respective prediction blocks subject to derivation.
  • a prediction block B0 adjacent to the top is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction blocks B0 and/or B1.
  • a prediction block B1 adjacent to the top is referred to, and a reference index of LX of a time merge candidate is set to the value of the reference index of LX of the prediction block B1. Since both prediction blocks B0 and B1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that prediction blocks A and B in the proximity do not perform LX prediction is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 21 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the first exemplary embodiment of the present embodiment.
  • PartMode is not N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) (NO in step S2101), that is, in case of 2N by 2N partition (PART_2Nx2N), 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), 2N by nD partition (PART_2NxnD), coding information of a prediction block A adjacent to the left is derived from the coding information storing memory 115 or 210 (step S2111).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) (YES in step S2101)
  • coding information of a prediction B adjacent to the top of the prediction block subject to derivation is derived from the coding information storing memory 115 or 210 (step S2117).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xB][yB] of LX of the prediction block B (step S2120)
  • xB and yB are indexes indicating the position of an upper left pixel in a prediction block B in a picture.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • Fig. 22 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the second exemplary embodiment of the present embodiment.
  • PartMode partition mode of a prediction block and the partition index PartIdx of the prediction block
  • PartIdx partition index of the prediction block
  • a prediction block adjacent to the edge outside of a coding block is referred to.
  • PartMode partition mode 2N by 2N partition (PART_2Nx2N)
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition(PART_2NxnU), or 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, as shown in Fig. 22B , 20C, and 20D , prediction blocks adjacent to the left edge of respective prediction blocks subject to derivation are referred to, and respective reference indexes of LX of time merge candidates are set to values of reference indexes of LX of prediction blocks adjacent to the left edge of respective prediction blocks subject to derivation.
  • a prediction block A0 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • a prediction block A1 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block A1. Since both prediction blocks A0 and A1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where coding block subject to processing is partitioned into two prediction blocks arranged horizontally, as shown in Fig. 22E, 22F, and 22G , a prediction block A0 adjacent to the left is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, reference indexes of LX of time merge candidates are set to a value of reference index of LX of prediction block A0.
  • a prediction block B1 adjacent to the top is referred to, and a reference index of LX of a time merge candidate is set to the value of the reference index of LX of the prediction block B1. Since both prediction blocks A0 and B1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that prediction blocks A and B in the proximity do not perform LX prediction is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 23 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the second exemplary embodiment of the present embodiment.
  • PartMode is not N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and the partition index PartIdx is not 1 (NO in step S2102), (i.e., in case of 2N by 2N partition (PART_2Nx2N), 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), 2N by nD partition (PART_2NxnD), N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), and the partition index PartIdx is not 1 (NO in step S210
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and the partition index PartIdx is 1 (YES in step S2102)
  • coding information of a prediction B adjacent to the top of the prediction block subject to derivation is derived from the coding information storing memory 115 or 210 (step S2117).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xB][yB] of LX of the prediction block B (step S2120).
  • xB and yB are indexes indicating the position of an upper left pixel in a prediction block B in a picture.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • Fig. 24 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the third exemplary embodiment of the present embodiment.
  • PartMode partition mode of a prediction block and the partition index PartIdx of the prediction block
  • which of prediction blocks: a prediction block adjacent to the left edge of a prediction block subject to derivation; or a prediction block adjacent to the upper edge thereof is used is switched.
  • a prediction block subject to derivation of which the partition index PartIdx is 0 a prediction block that is adjacent to a long edge, which has a higher correlation, is referred to.
  • a prediction block that is adjacent to an edge outside the prediction block is referred to.
  • a partition mode (PartMode) is 2N by 2N partition (PART_2Nx2N)
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), and 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, as shown in Fig. 24B, 24C, and 24D , a prediction block B0 adjacent to the upper edge, which is a long edge, is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, a reference index of LX of time merge candidate is set to a value of reference index of LX of prediction block B0.
  • a prediction block A1 adjacent to the left edge, which is outside the coding block is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block A1. Since both prediction blocks B0 and A1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where coding block subject to processing is partitioned into two prediction blocks arranged horizontally, as shown in Fig. 24E, 24F, and 24G , a prediction block A0 adjacent to the left edge, which is a long edge, is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, a reference index of LX of time merge candidate is set to a value of reference index of LX of prediction block A0.
  • a prediction block B1 adjacent to the upper edge, which is outside the coding block is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block B1. Since both prediction blocks A0 and B1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that prediction blocks A and B in the proximity do not perform LX prediction is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 25 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the third exemplary embodiment of the present embodiment.
  • PartMode partition mode
  • the partition mode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), 2N by nD partition (PART_2NxnD)
  • the partition index PartIdx is not 0, or in case that the partition mode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N)
  • the partition index PartIdx is not 1 (NO in step S2103), in other words, in case that the partition mode is 2N by 2N partition (PART_2Nx2N), in case that the partition mode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2Nxn
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and the partition index PartIdx is 1 (YES in step S2102)
  • coding information of a prediction B adjacent to the top of the prediction block subject to derivation is derived from the coding information storing memory 115 or 210 (step S2117).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xB][yB] of LX of the prediction block B (step S2120).
  • xB and yB are indexes indicating the position of an upper left pixel in a prediction block B in a picture.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • step S2119 through step S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • Fig. 26 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the fourth exemplary embodiment of the present embodiment.
  • PartMode partition mode of a coding block
  • PartIdx partition index of the prediction block
  • whether or not to refer to a prediction block adjacent to the left edge of a prediction block subject to derivation is switched.
  • the prediction block is referred to and in case that the prediction block adjacent to the left edge is inside the coding block, the prediction block is not referred to and a default value is set.
  • PartMode is 2N by 2N partition (PART_2Nx2N)
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), and 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, as shown in Fig. 26B, 26C, and 26D , a prediction block A0 adjacent to the left is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, and a reference index of LX of time merge candidate is set to a value of reference index of LX of prediction block A0.
  • a prediction block A1 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block A1. Since both prediction blocks A0 and A1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where coding block subject to processing is partitioned into two prediction blocks arranged horizontally, as shown in Fig. 26E, 26F, and 26G , a prediction block A0 adjacent to the left is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, reference indexes of LX of time merge candidates are set to a value of reference index of LX of prediction block A0.
  • a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0. Since a prediction block A0 to be referred to is outside of the coding block, a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that a prediction block A in the proximity does not perform LX prediction or in case that a partition index PartIdx of a prediction block subject to derivation is 1 is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 27 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the fourth exemplary embodiment of the present embodiment.
  • PartMode is not N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and the partition index PartIdx is not 1 (NO in step S2102), (i.e., in case of 2N by 2N partition (PART_2Nx2N), 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), 2N by nD partition (PART_2NxnD), N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), and the partition index PartIdx is not 1 (NO in step S210
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) and the partition index PartIdx is 1 (YES in step S2102)
  • subsequent process of step S2121 is performed for each L0 and L1 (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • Processes up to S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • Fig. 28 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the fifth exemplary embodiment of the present embodiment.
  • PartMode partition mode
  • a prediction block adjacent to the edge outside of a coding block is referred to.
  • the prediction block is referred to and in case that the prediction block adjacent to the left edge is inside the coding block, the prediction block is not referred to and a default value is set.
  • PartMode is 2N by 2N partition (PART_2Nx2N)
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), and 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, as shown in Fig. 28B, 28C, and 28D , a prediction block A0 adjacent to the left is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, a reference index of LX of time merge candidate is set to a value of reference index of LX of prediction block A0.
  • a prediction block A1 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to the value of reference index of LX of the prediction block A1. Since both prediction blocks A0 and A1 to refer to are outside of the coding block, a reference index of time merge candidate of two prediction blocks, of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • PartMode is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), where coding block subject to processing is partitioned into two prediction blocks arranged horizontally, in both cases of a prediction block of which the partition index PartIdx is 0 or 1, a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0.
  • a prediction block in the proximity of a prediction block subject to processing is not referred to, a prediction block in the coding block is not referred to, therefore a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that a prediction block A in the proximity does not perform LX prediction or in case that the partition mode (PartMode) of a coding block including a prediction block subject to derivation is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • PartMode partition mode of a coding block including a prediction block subject to derivation
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 29 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the fifth exemplary embodiment of the present embodiment.
  • PartMode is not N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N) (NO in step S2101), that is, in case of 2N by 2N partition (PART_2Nx2N), 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), 2N by nD partition (PART_2NxnD), coding information of a prediction block A adjacent to the left of a prediction block subject to derivation is derived from the coding information storing memory 115 or 210 (step S2111).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • step S2101 subsequent process of step S2121 is performed for each L0 and L1 (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • Processes up to S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • time merge candidates of two prediction blocks which is included in a same coding block and of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 without referring to a block in the proximity. Therefore, the derivation process can be simplified in comparison with the first, second, third, fourth, fifth, and seventh (described later) embodiment.
  • Fig. 30 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the sixth exemplary embodiment of the present embodiment.
  • a process of step S2115 is performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2115 is performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • Fig. 31 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the seventh exemplary embodiment of the present embodiment.
  • PartMode partition mode of a coding block including prediction mode subject to derivation (coding block subject to processing)
  • partition index PartIdx of a prediction mode subject to derivation prediction block A adjacent to the left of a coding block subject to processing is referred to as shown in Figs.
  • a reference index of LX of a time merge candidate is set to a value of a reference index of LX of the prediction block A. Since a prediction block A to be referred to is outside of the coding block, a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that the prediction block A in the proximity does not perform LX prediction is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • reference indexes of time merge candidates of respective prediction blocks in the coding block are set to a common value. Therefore, it is not required to derive a reference index of time merge candidate for each prediction block, so that the derivation process can be simplified.
  • Fig. 32 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the seventh exemplary embodiment of the present embodiment.
  • coding information on a prediction block A adjacent to the left of a coding block subject to processing is derived from the coding information storing memory 115 or 210 (step S2131).
  • step S2133 through step S2135 are performed for L0 and L1 respectively (step S2132 - S2136).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2134).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A, which is adjacent to the left of a coding block subject to processing, in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2135).
  • step S2133 through step S2135 are performed for L0 and L1 respectively (step S2132 - S2136), and the derivation process of a reference index is completed.
  • FIG. 43 shows adjacent blocks referred to in the derivation process of a reference index of a time merge candidate according to the eighth exemplary embodiment of the present embodiment.
  • PartMode partition mode of a coding block
  • PartIdx partition index of a prediction block
  • a prediction block adjacent to the left edge is referred to, and in case that the partition index PartIdx is a value other than 0, a prediction block in the proximity is not referred to, and a default value is set.
  • a prediction block adjacent to the left edge is always outside the coding block in any partition mode (PartMode).
  • PartMode partition mode
  • a prediction block adjacent to the left edge may be inside the coding block depending on the partition mode (PartMode).
  • a prediction block A0 adjacent to the left of a prediction block subject to derivation is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • PartMode is 2N by N partition (PART_2NxN), 2N by nU partition (PART_2NxnU), and 2N by nD partition (PART_2NxnD), where coding block subject to processing is partitioned into two prediction blocks arranged vertically, or in case that a partition mode (PartMode) is N by 2N partition (PART_Nx2N), nL by 2N partition (PART_nLx2N), or nR by 2N partition (PART_nRx2N), as shown in Fig.
  • a prediction block A0 adjacent to the left is referred to for a prediction block subject to derivation of which the partition index PartIdx is 0, and a reference index of LX of time merge candidate is set to a value of reference index of LX of prediction block A0.
  • a prediction block subject to derivation of which the partition index PartIdx is 1 a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0. Since a prediction block A0 to be referred to is outside of the coding block, a reference index of time merge candidate of two prediction blocks of which the partition index PartIdx are 0 and 1 respectively, can be derived in parallel, respectively.
  • a reference index of LX of a time merge candidate is set to a default value 0.
  • the reason why the default value of a reference index of LX of a time merge candidate is set to 0 in case that a prediction block A in the proximity does not perform LX prediction or in case that a partition index PartIdx of a prediction block subject to derivation is 1 is that a reference picture corresponding to the reference index value 0 has the highest possibility to be selected in inter prediction.
  • the scope of the present invention is not limited to this example.
  • the default value of the reference index may be set to a value other than 0 (e.g., 1, 2, etc.), or a syntax element indicating the default value of the reference index may be provided in a bitstream at a sequence level, at a picture level, or at a slice level, so that the syntax element can be transmitted and selected at encoding.
  • Fig. 44 shows a flowchart illustrating procedure of derivation process of a reference index of a time merge candidate in step S202 of Fig. 18 according to the eighth exemplary embodiment of the present embodiment.
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a value identical to the value of reference index refIdxLX[xA][yA] of LX of the prediction block A (step S2114).
  • xA and yA are indexes indicating the position of an upper left pixel in a prediction block A in a picture.
  • xN and yN are indexes indicating the position of an upper left pixel in a prediction block N in a picture.
  • the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0)
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 1
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 0.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N is 0, and the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction is 1.
  • the flag predFlagL0[xN][yN] indicating whether or not to use L0 prediction of the prediction block N
  • the flag predFlagL1[xN][yN] indicating whether or not to use L1 prediction are both 1.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2115).
  • step S2113 through step S2115 are performed for L0 and L1 respectively (step S2112 - S2116), and the derivation process of a reference index is completed.
  • step S2121 subsequent process of step S2121 is performed for each L0 and L1 (step S2118 - S2122).
  • LX is set to L0
  • LX is set to L1.
  • an inter prediction mode only includes L0 prediction (Pred_L0) and does not include L1 prediction (Pred_L1) nor bi-prediction (Pred_BI). Therefore a process regarding L1 can be omitted.
  • a reference index refIdxLXCol of LX of a time merge candidate is set to a default value 0 (step S2121).
  • Processes up to S2121 are performed for L0 and L1 respectively (step S2118 - S2122), and the derivation process of the reference index is completed.
  • N by N partition (PART_NxN) is not defined as a partition mode (PartMode) of a coding block according to the present embodiment
  • PartMode partition mode
  • a partition mode (PartMode) is N by N partition (PART_NxN)
  • a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0. Since a prediction block in the proximity of a prediction block subject to derivation is not referred to, reference indexes of time merge candidates of four prediction blocks of which the partition index PartIdx are 0, 1, 2, and 3 respectively, can be derived in parallel, respectively.
  • a reference index of a block in the proximity can be referred to so that a reference index of a time merge candidate can be derived.
  • Fig. 33 illustrates adjacent blocks referred to in the derivation process of a reference index of a time merge candidate of a coding block of N by N partition (PART_NxN).
  • a prediction block A0 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A0.
  • a prediction block B1 adjacent to the top is referred to, and a reference index of LX of a time merge candidate is set to the value of the reference index of LX of the prediction block B1.
  • a prediction block A2 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A2.
  • a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0.
  • prediction blocks A0 and A2 adjacent to the left of a prediction block subject to processing, and B1, adjacent to the top thereof, are all outside of the coding block
  • reference indexes of time merge candidates of four prediction blocks of which the partition index PartIdx are 0, 1, 2, and 3 respectively, can be derived in parallel, respectively.
  • a prediction block A0 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of the reference index of LX of the prediction block A0.
  • a prediction block A2 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A2.
  • a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0. Since both prediction blocks A0 and A2 adjacent to the left of a prediction block subject to processing are outside of the coding block, reference indexes of time merge candidates of four prediction blocks of which the partition index PartIdx are 0, 1, 2, and 3 can be derived respectively in parallel.
  • a prediction block A adjacent to the left of a coding block subject to processing is referred to, and a reference index of LX of a time merge candidate is set to a value of reference index of LX of the prediction block A. Since a prediction block A adjacent to the left of a coding block subject to processing is outside of the coding block, reference indexes of time merge candidates of four prediction blocks of which the partition index PartIdx are 0, 1, 2, and 3 can be derived respectively in parallel.
  • a prediction block A0 adjacent to the left is referred to, and a reference index of LX of a time merge candidate is set to a value of the reference index of LX of the prediction block A0.
  • a prediction block in the proximity is not referred to, and a reference index of LX of a time merge candidate is set to a default value 0. Since a prediction block A0 adjacent to the left of a prediction block subject to processing is outside of the coding block, reference indexes of time merge candidates of four prediction blocks of which the partition index PartIdx are 0, 1, 2, and 3 can be derived respectively in parallel.
  • Fig. 34 shows a flowchart illustrating procedure of time merge candidate derivation process of step S203 in Fig. 18 .
  • a picture colPic of a different time is derived by a slice type slice_type and a flag collocated_from_l0_flag described above (step S3101).
  • Fig. 35 shows a flowchart illustrating procedure of derivation process of a picture colPic of a different time of step S3101 in Fig. 34 .
  • a slice type slice_type is a B slice and a flag collocated_from_l0_flag described above is 0 (YES in step S3201 and YES in step S3202)
  • RefPicList1[0] i.e., a picture of which the reference index of reference list L1 is 0
  • step S3203 is defined as a picture colPic of a different time (step S3203).
  • RefPicList0[0] i.e., a picture of which the reference index of reference list L0 is 0
  • step S3205 RefPicList0[0]
  • a prediction block colPU of a different time is derived, and coding information is derived (step S3102).
  • Fig. 36 shows a flowchart illustrating procedure of derivation process of a prediction block colPU of a picture colPic of a different time of step S3102 in Fig. 34 .
  • a prediction block positioned at a lower right (outside) of a position identical to a prediction block subject to processing is used as a prediction block colPU of a different time (step S3301).
  • This prediction block corresponds to the prediction block T0 in Fig. 9 .
  • coding information of the prediction block colPU of a different time is derived (step S3302).
  • PredMode of the prediction block colPU of a different time cannot be used or in case that the prediction mode PredMode of the prediction block colPU of a different time is intra prediction (MODE_INTRA) (YES in step S3303 , and YES in step S3304)
  • a prediction block at an upper left in the middle of a position identical to a prediction block subject to processing is used as a prediction block colPU of a different time (step S3305).
  • This prediction block corresponds to the prediction block T1 in Fig. 9 .
  • a motion vector predictor mvLlCol of L0 that is derived from a prediction block in another picture at a position identical to a prediction block subject to encoding or decoding, and a flag availableFlagL0Col indicating whether or not a time merge candidate Col is valid are derived (step S3103), and a motion vector predictor mvLlCol of L1 and a flag availableFlagCol indicating whether or not the time merge candidate Col is valid are derived.
  • the flag availableFlagL0Col or the flag availableFlagLlCol is 1, the flag availableFlagCol indicating whether or not the time merge candidate Col is valid is set to 1.
  • Fig. 37 shows a flowchart illustrating procedure of derivation process of inter prediction information of time merge candidate of step S3103 and step 3104 in Fig. 34 .
  • LX a list of target to derive time merge candidate
  • LX prediction prediction using LX
  • step S3103 which is a derivation process of L0 of time merge candidate
  • step S3104 which is a derivation process of L1 of time merge candidate
  • LX is set to L1.
  • a prediction mode PredMode of a prediction block colPU of a different time is intra prediction (MODE_INTRA) or cannot be used, (NO in step S3401, NO in step S3402), a flag availableFlagLXCol and a flag predFlagLXCol are both set to 0 (step S3403), a motion vector mvLXCol is set to (0,0) (step S3404), the derivation process of inter prediction information of time merge candidate is completed.
  • mvCol, refIdxCol, and availableFlagCol are derived in the following procedure.
  • a motion vector mvCol is set to a value identical to that of MvL1[xPCol][yPCol], which is a motion vector of L1 of the prediction block colPU (step S3406)
  • a reference index refIdxCol is set to a value identical to that of a reference index RefIdxL1[xPCol][yPCol] of L1 (step S3407)
  • a list ListCol is set to L1 (step S3408).
  • xPCol and yPCol are indexes indicating the position of an upper left pixel of a prediction block colPU in a picture colPic of a different time.
  • a motion vector mvCol is set to a value identical to that of MvL0[xPCol][yPCol], which is a motion vector of L0 of the prediction block colPU (step S3410), and a reference index refIdxCol is set to a value identical to that of a reference index RefIdxL0[xPCol][yPCol] of L0 (step S3411), and a list ListCol is set to L0 (step S3412).
  • Fig. 38 shows a flowchart indicating procedure of derivation process of inter prediction information of time merge candidate when the inter prediction mode of the prediction block colPU is bi-prediction (Pred_BI).
  • step S3501 whether or not POCs of all pictures registered in all reference lists are smaller than POC of a picture subject to encoding or decoding currently is determined (step S3501).
  • POCs of all pictures registered in L0 or L1, L0 and L1 being all reference lists of prediction block colPU are smaller than POC of a picture subject to encoding or decoding currently (YES in step S3501) and LX is L0, that is, a prediction vector candidate of a motion vector of L0 of a picture subject to encoding or decode is derived (YES in step S3502), inter prediction information of L0 of the prediction block colPU is selected, and in case that LX is L1, that is, a prediction vector candidate of a motion vector of L1 of a picture subject to encoding or decode is derived (NO in step S3502), inter prediction information of L1 of the prediction block colPU is selected.
  • a motion vector mvCol is set to a value identical to that of MvL0[xPCol][yPCol] (step S3504), and a reference index refIdxCol is set to a value identical to that of RefIdxL0[xPCol][yPCol] (step S3505), and a list ListCol is set to L0 (step S3506).
  • a motion vector mvCol is set to a value identical to that of MvL1[xPCol][yPCol] (step S3507), and a reference index refIdxCol is set to a value identical to that of RefIdxL1[xPCol][yPCol] (step S3508), and a list ListCol is set to L1 (step S3509).
  • a flag availableFlagLXCol and a flag predFlagLXCol are both set to 1 (step S3414) .
  • a motion vector mvCol is scaled and defined as a motion vector mvLXCol of LX of a time merge candidate (step S3415).
  • An explanation will be given on procedure of scaling computation process of the motion vector, while referring to Figs. 39 and 40 .
  • Fig. 39 shows a flowchart indicating procedure of scaling computation process of the motion vector of step S3105 in Fig. 34 .
  • a distance td between pictures is derived (step S3601).
  • POC of a reference picture referred to in a list ListCol of a prediction block colPU is previous to the picture colPic of a different time in the display order
  • the distance td between pictures is a positive value
  • POC of a reference picture referred to in a list ListCol of a prediction block colPU is subsequent to the picture colPic of a different time in the display order
  • the distance td between pictures is a negative value.
  • tb POC of prediction block colPU of a different time ⁇ POC of a reference picture referred to in a list ListCol of a prediction block colPU .
  • a distance tb between pictures is derived (step S3602).
  • the distance tb between pictures is a positive value
  • the distance tb between pictures is a negative value
  • step S3603 distances td between pictures and tb between pictures are compared.
  • the motion vector mvLXCol of LX of time merge candidate is set to a value identical to that of the motion vector mvCol (step S3604), the scaling computation process is completed.
  • mvLXCol mvCol
  • step S3603 scaling computation process is performed by multiplying mvCol by tb or td in accordance with the following equation (step S3605), and a scaled motion vector mvLXCol of LX of time merge candidate is obtained.
  • mvLXCol tb or td * mvCol
  • Fig. 4 indicates an example in case that the scaling computation of step S3605 is performed with precision of integer. Processes of step S3606 - step S3608 in Fig. 40 correspond to the step S3605 of Fig. 40 .
  • a distance td between pictures and a distance tb between pictures are derived (step S3601 and step S3602).
  • step S3603 distances td between pictures and tb between pictures are compared.
  • the motion vector mvLXCol of LX of time merge candidate is set to a value identical to that of the motion vector mvCol (step S3604) in a similar manner to the flow chart in Fig. 39 , the scaling computation process is completed.
  • mvLXCol mvCol
  • tx 16384 + Abs td or 2 or td
  • DistScaleFactor tb * tx + 32 > > 6
  • mvLXCol ClipMv Sign DistScaleFactor * mvCol * Abs DistScaleFactor * mvCol + 127 > > 8
  • Fig. 41 shows a flowchart indicating a processing procedure for constructing a merge candidate list.
  • the method by setting priorities and registering a merge candidate from the higher priority in a merge candid list mergeCandList, the amount of code of merge index merge_idx[x0][y0] is reduced. By putting an element of higher priority in the upper part of the merge candidate list, the amount of code is reduced.
  • an index 0 of the merge candidate list is set to "0"
  • an index 1 thereof is set to "10”
  • an index 2 thereof is set to "110”
  • an index 3 thereof is set to "1110”
  • an index 4 thereof is set to "11110"
  • the merge candidate list mergeCandList is configured as a list structure, and in which a merge index indicating locations in the merge candidate list, and a storage area that stores a merge candidate corresponding to the index as an element are provided.
  • the number of merge index starts from 0, and a merge candidate is stored in a storage area of the merge candidate list mergeCandList.
  • mergeCandList[i] a prediction block that becomes a merge candidate of a mergeindex i registered in the merge candidate list mergeCandList.
  • a merge candidate A is registered at the top of the merge candidate list mergeCandList (step S4102).
  • a merge candidate B is registered at the end of the merge candidate list mergeCandList (step S4104).
  • a merge candidate C is registered at the end of the merge candidate list mergeCandList (step S4106).
  • a merge candidate D is registered at the end of the merge candidate list mergeCandList (step S4108).
  • step S4109 a merge candidate E is registered at the end of the merge candidate list mergeCandList (step S4110).
  • vailableFlagCol is 1 (YES in step S4109)
  • a merge candidate Col is registered at the end of the merge candidate list mergeCandList (step S4110).
  • a prediction block A adjacent to the left and a prediction block B adjacent to the top often with the prediction block subject to encoding or decoding in an integrated manner.
  • the merge candidates A and B are registered ahead of other merge candidates C, D, E, and Col in the candidate list, preferentially.
  • a merge candidate is selected from merge candidates registered in the merge candidate list, and a merge index and inter prediction information of a merge candidate corresponding to the merge index is provided to the motion compensation prediction unit 105.
  • a method similar to that of the prediction method determining unit 107 can be used. Coding information and coding distortion between an amount of code of coding information and a residual signal, and coding distortion between a prediction picture signal and a picture signal are derived for each merge candidate. A merge candidate that has the least amount of generated code and coding distortion is determined. For each merge candidate, entropy encoding is performed of syntax element merge_idx, which is coding information of a merge mode, is performed, and an amount of code of coding information is calculated. Further, for each merge candidate, an amount of code of an encoded prediction residual signal is calculated with a method similar to that of the motion compensation prediction unit 105.
  • the prediction residual signal is a residual signal between a prediction picture signal, on which motion compensation is performed in accordance with inter prediction information of each merge candidate, and a picture signal, which is subject to encoding and which is provided from the picture memory 101.
  • Total amount of code to which an amount of code of coding information (i.e., a merge index) and an amount of code of prediction residual signal is added is calculated and set as an evaluation value.
  • a coding distortion is calculated as a ratio representing an error from an original picture signal, the error arising from encoding.
  • coding information of less amount of generated code and less coding distortion is determined.
  • a merge index corresponding to the determined coding information is encoded as a flag merge_idx represented by a second syntax pattern of a prediction block.
  • the amount of generated code calculated in this process is preferably calculated by simulating coding processes. However, approximations to simplify calculations or a rough estimate can also be applied.
  • a merge candidate corresponding to a provided merge index is selected from merge candidates registered in the merge candidate list, and inter prediction information of the merge candidate is provided to the motion compensation prediction unit 206, and stored in the coding information storing memory 210.
  • a bitstream of moving pictures outputted from the moving picture encoding device according to the embodiments described above has a specific data format so that the data stream can be decoded in accordance with the coding method used in the embodiments.
  • a moving picture decoding device corresponding to the moving picture encoding device can decode the bitstream in this specific format.
  • a bitstream may be transformed into a data format appropriate to the manner of transmission on the communication path so as to be transmitted.
  • a moving picture transmission device and a moving picture receiver device are provided.
  • the moving picture transmission device transforms a bitstream output from the moving picture encoding device into encoded data in a data format appropriate to the manner of transmission via the communication path and transmits the encoded data to the network.
  • the moving picture receiver device receives encoded data from the network, reconstructs a bitstream, and provides the bitstream to the moving picture decoding device.
  • the moving picture transmission device includes a memory that buffers a bitstream output from the moving picture encoding device, a packet processing unit that packetizes the bitstream, a transmission unit that transmits the packetized bitstream via a network.
  • the moving picture receiver device includes a receiving unit that receives a packetized bitstream via a network, a memory that buffers the received bitstream, and a packet processing unit that processes a packet of the encoded data so as to generate a bitstream, and provides the bitstream to the moving picture decoding device.
  • the processes relating to encoding and decoding described above can be implemented as a transmission device, a storage, and/or a receiver device using hardware.
  • the process can also be implemented by firmware stored in a read only memory (ROM), a flash memory, or the like, or by software of computers or the like.
  • the firmware program or the software program can be stored in a recording medium readable by a computer or the like and provided, can be provided from a server via a wired or wireless network, or can be provided by data broadcasting as terrestrial broadcasting or satellite digital broadcasting.
  • the present invention is applicable to a moving picture encoding technology utilizing motion compensation prediction.
EP21181798.6A 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung Pending EP3907997A1 (de)

Applications Claiming Priority (9)

Application Number Priority Date Filing Date Title
JP2011276457 2011-12-16
JP2011276456 2011-12-16
JP2011289287 2011-12-28
JP2011289288 2011-12-28
PCT/JP2012/008016 WO2013088738A1 (ja) 2011-12-16 2012-12-14 動画像符号化装置、動画像符号化方法及び動画像符号化プログラム、並びに動画像復号装置、動画像復号方法及び動画像復号プログラム
EP19217766.5A EP3654651B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP19156821.1A EP3503554B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung
EP20169995.6A EP3703371B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP12857437.3A EP2802146B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung

Related Parent Applications (5)

Application Number Title Priority Date Filing Date
EP12857437.3A Division EP2802146B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP20169995.6A Division EP3703371B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP20169995.6A Division-Into EP3703371B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP19156821.1A Division EP3503554B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung
EP19217766.5A Division EP3654651B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung

Publications (1)

Publication Number Publication Date
EP3907997A1 true EP3907997A1 (de) 2021-11-10

Family

ID=49479186

Family Applications (6)

Application Number Title Priority Date Filing Date
EP19217766.5A Active EP3654651B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP20169995.6A Active EP3703371B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP12857437.3A Active EP2802146B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP19156821.1A Active EP3503554B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung
EP20169994.9A Active EP3703370B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP21181798.6A Pending EP3907997A1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung

Family Applications Before (5)

Application Number Title Priority Date Filing Date
EP19217766.5A Active EP3654651B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP20169995.6A Active EP3703371B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP12857437.3A Active EP2802146B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
EP19156821.1A Active EP3503554B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung
EP20169994.9A Active EP3703370B1 (de) 2011-12-16 2012-12-14 Vorrichtung zur dynamischen bildcodierung, verfahren zur dynamischen bildcodierung, programm zur dynamischen bildcodierung, vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung

Country Status (12)

Country Link
US (5) US9674545B2 (de)
EP (6) EP3654651B1 (de)
BR (1) BR112014013962B1 (de)
DK (5) DK3654651T3 (de)
ES (3) ES2877369T3 (de)
HR (1) HRP20211359T1 (de)
HU (5) HUE043876T2 (de)
PL (5) PL3703371T3 (de)
RU (4) RU2598302C2 (de)
SI (5) SI3703370T1 (de)
TW (5) TWI586153B (de)
WO (1) WO2013088738A1 (de)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
ES2705251B1 (es) * 2011-09-23 2020-01-29 Kt Corp Método para inducir un bloque candidato de fusión y dispositivo que usa el mismo
WO2016068685A1 (ko) * 2014-10-31 2016-05-06 삼성전자 주식회사 고정밀 스킵 부호화를 이용한 비디오 부호화 장치 및 비디오 복호화 장치 및 그 방법
US10477233B2 (en) * 2015-09-30 2019-11-12 Apple Inc. Predictor candidates for motion estimation search systems and methods
KR102275420B1 (ko) * 2016-07-12 2021-07-09 한국전자통신연구원 영상 부호화/복호화 방법 및 이를 위한 기록 매체
CN116506597A (zh) * 2016-08-03 2023-07-28 株式会社Kt 视频解码方法、视频编码方法和视频数据的传送方法
US11470346B2 (en) * 2018-05-09 2022-10-11 Sharp Kabushiki Kaisha Systems and methods for performing motion vector prediction using a derived set of motion vectors
US10855986B2 (en) 2018-05-29 2020-12-01 Qualcomm Incorporated Bandwidth compression for neural network systems
IL291768A (en) * 2019-10-03 2022-06-01 Huawei Tech Co Ltd Encoding process for geometric division mode
CN111263151B (zh) * 2020-04-26 2020-08-25 腾讯科技(深圳)有限公司 视频编码方法、装置、电子设备和计算机可读存储介质
US11949863B2 (en) * 2021-08-02 2024-04-02 Tencent America LLC Geometric partition mode with intra block copy
KR20240054296A (ko) * 2021-08-18 2024-04-25 엘지전자 주식회사 채널간 상관도에 기반한 피쳐 부호화/복호화 방법, 장치 및 비트스트림을 저장한 기록 매체

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1406453A1 (de) * 2002-10-04 2004-04-07 Lg Electronics Inc. Direkt-Modus Bewegungsvektorberechnung für B-Bilder

Family Cites Families (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3968712B2 (ja) * 2003-04-28 2007-08-29 ソニー株式会社 動き予測補償装置及びその方法
US7145950B2 (en) * 2003-07-14 2006-12-05 Primax Electronics Ltd. Method of motion vector determination in digital video compression
JP4591657B2 (ja) * 2003-12-22 2010-12-01 キヤノン株式会社 動画像符号化装置及びその制御方法、プログラム
KR101441269B1 (ko) * 2005-09-26 2014-09-18 미쓰비시덴키 가부시키가이샤 동화상 복호 장치 및 동화상 복호 방법
CN100525446C (zh) * 2005-12-26 2009-08-05 凌阳科技股份有限公司 可纵向译码输出的区块译码方法及装置
JP4790446B2 (ja) * 2006-03-01 2011-10-12 三菱電機株式会社 動画像復号装置及び動画像符号化装置
EP2079242A4 (de) * 2006-10-30 2010-11-03 Nippon Telegraph & Telephone Verfahren zur erzeugung prädiktiver referenzinformationen, verfahren zur dynamischen bildkodierung und -dekodierung, vorrichtung dafür, programm dafür und speichermedium mit dem programm
US7991237B2 (en) * 2007-06-28 2011-08-02 Mitsubishi Electric Corporation Image encoding device, image decoding device, image encoding method and image decoding method
JP2009164880A (ja) * 2008-01-07 2009-07-23 Mitsubishi Electric Corp トランスコーダ及び受信機
JP5406465B2 (ja) * 2008-04-24 2014-02-05 株式会社Nttドコモ 画像予測符号化装置、画像予測符号化方法、画像予測符号化プログラム、画像予測復号装置、画像予測復号方法及び画像予測復号プログラム
JP5401071B2 (ja) * 2008-10-09 2014-01-29 株式会社Nttドコモ 動画像符号化装置、動画像復号装置、動画像符号化方法、動画像復号方法、動画像符号化プログラム、動画像復号プログラム、動画像処理システムおよび動画像処理方法
EP2237557A1 (de) * 2009-04-03 2010-10-06 Panasonic Corporation Kodierung für Filterkoeffizienten
KR101456498B1 (ko) * 2009-08-14 2014-10-31 삼성전자주식회사 계층적 부호화 단위의 스캔 순서를 고려한 비디오 부호화 방법 및 장치, 비디오 복호화 방법 및 장치
US8971400B2 (en) * 2010-04-14 2015-03-03 Mediatek Inc. Method for performing hybrid multihypothesis prediction during video coding of a coding unit, and associated apparatus
US10171813B2 (en) 2011-02-24 2019-01-01 Qualcomm Incorporated Hierarchy of motion prediction video blocks
US9866859B2 (en) * 2011-06-14 2018-01-09 Texas Instruments Incorporated Inter-prediction candidate index coding independent of inter-prediction candidate list construction in video coding
US9313494B2 (en) * 2011-06-20 2016-04-12 Qualcomm Incorporated Parallelization friendly merge candidates for video coding
KR20130050406A (ko) 2011-11-07 2013-05-16 오수미 머지 모드에서의 움직임 정보 생성 방법
KR20130050407A (ko) 2011-11-07 2013-05-16 오수미 인터 모드에서의 움직임 정보 생성 방법
JP5561348B2 (ja) 2011-12-16 2014-07-30 株式会社Jvcケンウッド 動画像復号装置、動画像復号方法、及び動画像復号プログラム、並びに、受信装置、受信方法、及び受信プログラム

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1406453A1 (de) * 2002-10-04 2004-04-07 Lg Electronics Inc. Direkt-Modus Bewegungsvektorberechnung für B-Bilder
JP2004129191A (ja) 2002-10-04 2004-04-22 Lg Electronics Inc Bピクチャのダイレクトモード動きベクトル演算方法

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
I-K KIM ET AL: "CE2: Test result of asymmetric motion partition (AMP)", 6. JCT-VC MEETING; 97. MPEG MEETING; 14-7-2011 - 22-7-2011; TORINO; (JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11 AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-F379, 2 July 2011 (2011-07-02), XP030009402 *
J-L LIN ET AL: "Improved Advanced Motion Vector Prediction", 4. JCT-VC MEETING; 95. MPEG MEETING; 20-1-2011 - 28-1-2011; DAEGU;(JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-D125, 15 January 2011 (2011-01-15), XP030008165, ISSN: 0000-0015 *
NAKAMURA H ET AL: "Non-CE9: Construction of MVP list without using scaling", 98. MPEG MEETING; 28-11-2011 - 2-12-2011; GENEVA; (MOTION PICTURE EXPERT GROUP OR ISO/IEC JTC1/SC29/WG11),, no. m21778, 23 November 2011 (2011-11-23), XP030050341 *
NAKAMURA H ET AL: "Non-CE9: Derivation process of reference indices for temporal merging candidates", 8. JCT-VC MEETING; 99. MPEG MEETING; 1-2-2012 - 10-2-2012; SAN JOSE; (JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11 AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-H0199, 20 January 2012 (2012-01-20), XP030111226 *
ZHANG (HUAWEI) L ET AL: "Modification for robust parsing of AMVP", 5. JCT-VC MEETING; 96. MPEG MEETING; 16-3-2011 - 23-3-2011; GENEVA;(JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-E302, 11 March 2011 (2011-03-11), XP030008808, ISSN: 0000-0005 *
ZHENG Y ET AL: "Unified Motion Vector Predictor Selection for Merge and AMVP", 5. JCT-VC MEETING; 96. MPEG MEETING; 16-3-2011 - 23-3-2011; GENEVA;(JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-E396, 11 March 2011 (2011-03-11), XP030008902, ISSN: 0000-0005 *

Also Published As

Publication number Publication date
SI3654651T1 (sl) 2021-08-31
EP2802146B1 (de) 2019-02-27
US20140286414A1 (en) 2014-09-25
US9743104B2 (en) 2017-08-22
PL3703370T3 (pl) 2021-12-06
RU2598302C2 (ru) 2016-09-20
US9674545B2 (en) 2017-06-06
US20170238009A1 (en) 2017-08-17
US20160309180A1 (en) 2016-10-20
TWI586153B (zh) 2017-06-01
TW201332375A (zh) 2013-08-01
RU2014129060A (ru) 2016-02-10
EP3703371A1 (de) 2020-09-02
SI3703371T1 (sl) 2021-11-30
EP3654651B1 (de) 2021-04-21
DK3654651T3 (da) 2021-05-31
RU2678507C9 (ru) 2020-11-12
HRP20211359T1 (hr) 2021-11-26
EP3703370A1 (de) 2020-09-02
TWI593277B (zh) 2017-07-21
ES2877369T3 (es) 2021-11-16
US20170238010A1 (en) 2017-08-17
EP3503554B1 (de) 2020-02-05
EP2802146A1 (de) 2014-11-12
SI3703370T1 (sl) 2021-11-30
BR112014013962B1 (pt) 2020-04-07
ES2888805T3 (es) 2022-01-07
DK3703370T3 (da) 2021-09-06
PL3654651T3 (pl) 2021-09-27
EP3654651A1 (de) 2020-05-20
TW201642659A (zh) 2016-12-01
SI2802146T1 (sl) 2019-07-31
RU2679552C1 (ru) 2019-02-11
US20170238008A1 (en) 2017-08-17
EP3503554A1 (de) 2019-06-26
PL3503554T3 (pl) 2020-06-29
TW201720165A (zh) 2017-06-01
PL2802146T3 (pl) 2019-07-31
TWI578766B (zh) 2017-04-11
HUE047995T2 (hu) 2020-05-28
TWI607649B (zh) 2017-12-01
HUE055991T2 (hu) 2022-01-28
RU2678507C1 (ru) 2019-01-29
SI3503554T1 (sl) 2020-06-30
TWI559746B (zh) 2016-11-21
US9906810B2 (en) 2018-02-27
PL3703371T3 (pl) 2021-12-06
BR112014013962A2 (pt) 2017-06-13
DK3703371T3 (da) 2021-09-06
TW201720163A (zh) 2017-06-01
TW201720164A (zh) 2017-06-01
EP2802146A4 (de) 2015-09-23
US9906809B2 (en) 2018-02-27
HUE043876T2 (hu) 2019-09-30
DK3503554T3 (da) 2020-05-11
HUE053940T2 (hu) 2021-08-30
EP3703370B1 (de) 2021-08-04
BR112014013962A8 (pt) 2017-06-13
WO2013088738A1 (ja) 2013-06-20
ES2886342T3 (es) 2021-12-17
DK2802146T3 (en) 2019-04-15
US9906811B2 (en) 2018-02-27
HUE055993T2 (hu) 2022-01-28
EP3703371B1 (de) 2021-08-04
RU2636688C1 (ru) 2017-11-27

Similar Documents

Publication Publication Date Title
EP2802146B1 (de) Vorrichtung zur dynamischen bilddecodierung, verfahren zur dynamischen bilddecodierung und programm zur dynamischen bilddecodierung
JP5942782B2 (ja) 動画像復号装置、動画像復号方法、動画像復号プログラム、受信装置、受信方法及び受信プログラム
JP6229806B2 (ja) 動画像復号装置、動画像復号方法、動画像復号プログラム、受信装置、受信方法及び受信プログラム
JP6644270B2 (ja) 動画像符号化装置、動画像符号化方法、動画像符号化プログラム、送信装置、送信方法及び送信プログラム
RU2666275C1 (ru) Устройство и способ кодирования движущегося изображения, долговременный считываемый компьютером носитель записи, на который записана программа кодирования изображения

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20210625

AC Divisional application: reference to earlier application

Ref document number: 2802146

Country of ref document: EP

Kind code of ref document: P

Ref document number: 3503554

Country of ref document: EP

Kind code of ref document: P

Ref document number: 3654651

Country of ref document: EP

Kind code of ref document: P

Ref document number: 3703371

Country of ref document: EP

Kind code of ref document: P

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

B565 Issuance of search results under rule 164(2) epc

Effective date: 20210806

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: EXAMINATION IS IN PROGRESS

17Q First examination report despatched

Effective date: 20230102